Tonal impact on diphthong realizations in Standard Mandarin Chinese: GAMMs Modelling codes

1 Preparations before modelling

1.1 Import packages for modelling

library(tidyverse)
library(nlme)
library(ggplot2)
library(scales)
library(ggpubr)
library(mgcv)
library(itsadug)
# for Lobanov transformation
library(phonR)

1.2 Set the number of cores for computation

ncores<-18

1.3 Change the color plan

library(colorspace)
library(scales)
# heatmap like
mapcols  <- rev(colorRamps::matlab.like2(100))
# a "more transparent" heatmap like
mapcols_pastel <- lighten(desaturate(mapcols, amount = 0.3), amount = 0.2)

1.4 Change the range of x-axis in the plotting

# This is for normalized time scale
ticknames<-c(0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1)
tickvals <- c(0,1,2,3,4,5,6,7,8,9,10)
# For re-transformed real-time scale
ticknames1<-c(0, 20, 40, 60, 80, 100, 120, 140, 160)
tickvals1 <- c(0, 20, 40, 60, 80, 100, 120, 140, 160)
ticknames2<-c(0, 25, 50, 75, 100, 125, 150, 175, 200)
tickvals2<- c(0, 25, 50, 75, 100, 125, 150, 175, 200)

1.5 import dataframes

The data used in the analysis are obtained from the corpus AISHELL-1 published by Beijing Shell Shell Technology Co.,Ltd. The recordings are conducted in a quiet indoor environment using a high fidelity microphone and downsampled to 16kHz. The sample used in the study are from 10 male speakers and 10 female speakers of Standard Mandarin. We obtained 4 sub-datasets, of two falling diphthongs /ai/ and /au/ by two genders.

load("~/dataaimas.Rda")

load("~/dataaifem.Rda")

load("~/dataaumas.Rda")

load("~/dataaufem.Rda")

2 Overview of the dataset

head(data.ai.fem)

unique(au.fem.bis$leftSegment)

 [1] "t"  "k"  "p"  "ʂ"  "ʈʂ" "m"  "n"  "ts" "l"  "ʐ"  "ʔ"  "s" 

numero is the token number when extracting the data; sex is the gender of the speaker; speaker is the speaker ID; file is the audio file where the extracted token from; measurement.no is the normalized time point from 0 to 10; diphthong is what the diphthong is; tone is the tone of this token (tone 3 sandhi is transformed to tone 2); toneOri is the original tonal category; tone.ord is the tone as the ordered factor value (1<2<3<4); tonebis.ord is the reorganized ordered factor value (4<1<2<3) (see Appendix B); word is the word which the syllable belongs to; leftTone is the pre-target tone; rightTone is the post-target tone; initial is whether the character is at the beginning of the sentence (in the current study, the value should always be N); aspiration is whether the segment at onset is aspirated (in this study, the value should always be no); duration..ms. is the duration of the diphthong in ms; f1, f2, f0 are F1, F2, f0 in Hz; durationZscore2, f1Zscore2, f2Zscore2, f0Zscore2 are the values in Z-score; start is whether the measurement point is at the start point of all the 11 points (this is for the auto-regressive modelling); posR is the relative position in the sentence; wordPos, speakerPos, wordLeftRightTone, speakerLeftRightTone are word information and speaker information adjusted by posR and LeftTone and RightTone.

names(data.ai.fem)

 [1] "numero"               "sex"                  "speaker"              "file"                 "measurement.no"      
 [6] "diphthong"            "tone"                 "toneOri"              "tone.ord"             "toneBis.ord"         
[11] "word"                 "leftTone"             "rightTone"            "initial"              "aspiration"          
[16] "duration..ms."        "f1"                   "f2"                   "f0"                   "durationZscore2"     
[21] "f1Zscore2"            "f2Zscore2"            "f0Zscore2"            "start"                "posR"                
[26] "wordPos"              "speakerPos"           "speakerLeftRightTone" "wordLeftRightTone"   

3 For plotting: reconstruction of the data to the absolute scale

3.1 /ai/ male

meanDurationZByTone <- aggregate(data.ai.mas$durationZscore2, by = list(tone = data.ai.mas$tone.ord),mean)

# global mean values and standard deviations
## f0
global_mean0 <- mean(data.ai.mas$f0, na.rm = TRUE)
global_sd0 <- sd(data.ai.mas$f0, na.rm = TRUE)
## f1
global_mean1 <- mean(data.ai.mas$f1, na.rm = TRUE)
global_sd1 <- sd(data.ai.mas$f1, na.rm = TRUE)
## f2
global_mean2 <- mean(data.ai.mas$f2, na.rm = TRUE)
global_sd2 <- sd(data.ai.mas$f2, na.rm = TRUE)
## duration
global_meand <- mean(data.ai.mas$duration..ms., na.rm = TRUE)
global_sdd <- sd(data.ai.mas$duration..ms., na.rm = TRUE)
# print results
print(global_mean0)

[1] 130.9548

print(global_sd0)

[1] 23.29733

print(global_mean1)

[1] 621.0464

print(global_sd1)

[1] 107.4572

print(global_mean2)

[1] 1684.322

print(global_sd2)

[1] 152.3606

print(global_meand)

[1] 139.8406

print(global_sdd)

[1] 40.63055

# reconstructed absolute duration of 4 tones
durationT1 <- meanDurationZByTone[1,2] * global_sdd + global_meand
durationT2 <- meanDurationZByTone[2,2] * global_sdd + global_meand
durationT3 <- meanDurationZByTone[3,2] * global_sdd + global_meand
durationT4 <- meanDurationZByTone[4,2] * global_sdd + global_meand
print(durationT1)

[1] 148.2416

print(durationT2)

[1] 138.5754

print(durationT3)

[1] 137.0176

print(durationT4)

[1] 140.0018

3.2 /ai/ female

meanDurationZByTonef <- aggregate(data.ai.fem$durationZscore2, by = list(tone = data.ai.fem$tone.ord),mean)

# global mean values and standard deviations
## f0
global_mean0f <- mean(data.ai.fem$f0, na.rm = TRUE)
global_sd0f <- sd(data.ai.fem$f0, na.rm = TRUE)
## f1
global_mean1f <- mean(data.ai.fem$f1, na.rm = TRUE)
global_sd1f <- sd(data.ai.fem$f1, na.rm = TRUE)
## f2
global_mean2f <- mean(data.ai.fem$f2, na.rm = TRUE)
global_sd2f <- sd(data.ai.fem$f2, na.rm = TRUE)
## duration
global_meandf <- mean(data.ai.fem$duration..ms., na.rm = TRUE)
global_sddf <- sd(data.ai.fem$duration..ms., na.rm = TRUE)
# print results
print(global_mean0f)

[1] 219.3686

print(global_sd0f)

[1] 33.53907

print(global_mean1f)

[1] 766.5056

print(global_sd1f)

[1] 157.2238

print(global_mean2f)

[1] 1896.827

print(global_sd2f)

[1] 218.7678

print(global_meandf)

[1] 152.4137

print(global_sddf)

[1] 32.45953

# reconstructed absolute duration of 4 tones
durationT1f <- meanDurationZByTonef[1,2] * global_sddf + global_meandf
durationT2f <- meanDurationZByTonef[2,2] * global_sddf + global_meandf
durationT3f <- meanDurationZByTonef[3,2] * global_sddf + global_meandf
durationT4f <- meanDurationZByTonef[4,2] * global_sddf + global_meandf
print(durationT1f)

[1] 144.0195

print(durationT2f)

[1] 156.5566

print(durationT3f)

[1] 152.7313

print(durationT4f)

[1] 151.7184

##/au/ male

meanDurationZByToneau <- aggregate(data.au.mas$durationZscore2, by = list(tone = data.au.mas$tone.ord),mean)

# global mean values and standard deviations
## f0
global_mean0au <- mean(data.au.mas$f0, na.rm = TRUE)
global_sd0au <- sd(data.au.mas$f0, na.rm = TRUE)
## f1
global_mean1au <- mean(data.au.mas$f1, na.rm = TRUE)
global_sd1au <- sd(data.au.mas$f1, na.rm = TRUE)
## f2
global_mean2au <- mean(data.au.mas$f2, na.rm = TRUE)
global_sd2au <- sd(data.au.mas$f2, na.rm = TRUE)
## duration
global_meandau <- mean(data.au.mas$duration..ms., na.rm = TRUE)
global_sddau <- sd(data.au.mas$duration..ms., na.rm = TRUE)

# print results
print(global_mean0au)

[1] 129.9584

print(global_sd0au)

[1] 24.0711

print(global_mean1au)

[1] 609.4632

print(global_sd1au)

[1] 94.02331

print(global_mean2au)

[1] 1175.144

print(global_sd2au)

[1] 171.9271

print(global_meandau)

[1] 138.8545

print(global_sddau)

[1] 35.32614

# reconstructed absolute duration of 4 tones
durationT1au <- meanDurationZByToneau[1,2] * global_sddau + global_meandau
durationT2au <- meanDurationZByToneau[2,2] * global_sddau + global_meandau
durationT3au <- meanDurationZByToneau[3,2] * global_sddau + global_meandau
durationT4au <- meanDurationZByToneau[4,2] * global_sddau + global_meandau
print(durationT1au)

[1] 149.8533

print(durationT2au)

[1] 138.4197

print(durationT3au)

[1] 130.7606

print(durationT4au)

[1] 139.0956

3.3 /au/ female

meanDurationZByToneauf <- aggregate(data.au.fem$durationZscore2, by = list(tone = data.au.fem$tone.ord),mean)

# global mean values and standard deviations
## f0
global_mean0auf <- mean(data.au.fem$f0, na.rm = TRUE)
global_sd0auf <- sd(data.au.fem$f0, na.rm = TRUE)
## f1
global_mean1auf <- mean(data.au.fem$f1, na.rm = TRUE)
global_sd1auf <- sd(data.au.fem$f1, na.rm = TRUE)
## f2
global_mean2auf <- mean(data.au.fem$f2, na.rm = TRUE)
global_sd2auf <- sd(data.au.fem$f2, na.rm = TRUE)
## duration
global_meandauf <- mean(data.au.fem$duration..ms., na.rm = TRUE)
global_sddauf <- sd(data.au.fem$duration..ms., na.rm = TRUE)

# print results
print(global_mean0auf)

[1] 223.0908

print(global_sd0auf)

[1] 33.08992

print(global_mean1auf)

[1] 751.3295

print(global_sd1auf)

[1] 127.0412

print(global_mean2auf)

[1] 1348.953

print(global_sd2auf)

[1] 191.1671

print(global_meandauf)

[1] 158.5503

print(global_sddauf)

[1] 30.91023

# reconstructed absolute duration of 4 tones
durationT1auf <- meanDurationZByToneauf[1,2] * global_sddauf + global_meandauf
durationT2auf <- meanDurationZByToneauf[2,2] * global_sddauf + global_meandauf
durationT3auf <- meanDurationZByToneauf[3,2] * global_sddauf + global_meandauf
durationT4auf <- meanDurationZByToneauf[4,2] * global_sddauf + global_meandauf
print(durationT1auf)

[1] 159.5236

print(durationT2auf)

[1] 163.6863

print(durationT3auf)

[1] 153.0105

print(durationT4auf)

[1] 159.554

4 Model 1: Diphthong realization based on tones

4.1 Model 1A: /ai/ male, F1 as output

gamm.model1a.noAR <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  

                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1a.noAR, paste("Gamm_model1a_noAR.rds"))

gamm.model1a.noAR <- 
  readRDS("Gamm_model1a_noAR.rds")

r.gamm.model1a <- start_value_rho(gamm.model1a.noAR)

# Auto-regressive model

gamm.model1a <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.ai.mas, method="fREML", rho = r.gamm.model1a, AR.start = data.ai.mas$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1a, paste("Gamm_model1a.rds"))

gamm.model1a <- 
  readRDS("Gamm_model1a.rds")

summary(gamm.model1a, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept)   0.13798    0.04700   2.936  0.00334 **
toneBis.ord1 -0.43108    0.16979  -2.539  0.01113 * 
toneBis.ord2  0.08773    0.09831   0.892  0.37223   
toneBis.ord3  0.03179    0.11072   0.287  0.77402   
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df      F p-value    
s(measurement.no)              7.805  8.046 69.293  <2e-16 ***
s(measurement.no):toneBis.ord1 2.093  2.662  0.741  0.5883    
s(measurement.no):toneBis.ord2 7.016  7.863  9.458  <2e-16 ***
s(measurement.no):toneBis.ord3 3.813  4.531  2.577  0.0272 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =   0.64   Deviance explained = 67.4%
fREML = 9248.7  Scale est. = 0.30499   n = 12023

gam.check(gamm.model1a)

Method: fREML   Optimizer: perf chol
$grad
 [1]  4.041212e-14 -2.884692e-12 -9.547918e-14 -2.406964e-13 -1.030642e-11 -5.341571e-05 -9.061204e-05 -9.884080e-05 -4.978977e-05
[10]  7.482903e-14 -5.014362e-05 -4.804637e-05 -1.836753e-11  6.714629e-13 -8.126086e-05 -1.823430e-12 -1.267253e-11 -1.749711e-13
[19]  6.743051e-12  1.731948e-13 -5.776357e-11 -2.247091e-13 -4.232951e-05 -1.195044e-12 -4.256151e-12 -3.446132e-13 -4.647838e-12
[28]  2.144951e-13 -9.968915e-11  5.329071e-14 -8.219914e-11  4.916956e-12 -1.794973e-10 -6.252776e-13  1.289663e-09

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.514578e+00  1.199872e-02  8.078218e-02  2.060962e-02 -8.069430e-02  2.692505e-09  5.347152e-07  1.793837e-09  1.713211e-07
   1.199872e-02  2.713252e-01 -3.028790e-03 -1.759066e-02  7.782067e-03  6.629533e-09  8.576933e-06  2.411988e-08 -3.135409e-08
   8.078218e-02 -3.028790e-03  2.322656e+00 -2.659205e-02  2.079923e-02 -4.780141e-09 -5.271237e-07  1.432835e-09  6.679834e-06
   2.060962e-02 -1.759066e-02 -2.659205e-02  8.191930e-01  1.576916e-02 -1.611386e-08 -8.931354e-07 -1.034232e-09  7.662352e-08
  -8.069430e-02  7.782067e-03  2.079923e-02  1.576916e-02  2.209977e+01 -1.408603e-07 -1.831717e-05 -9.465974e-08 -3.547565e-05
   2.692505e-09  6.629533e-09 -4.780141e-09 -1.611386e-08 -1.408603e-07  5.341514e-05  7.537970e-12 -6.838703e-11  5.372229e-12
   5.347152e-07  8.576933e-06 -5.271237e-07 -8.931354e-07 -1.831717e-05  7.537970e-12  9.063017e-05  9.596475e-10 -8.123521e-11
   1.793837e-09  2.411988e-08  1.432835e-09 -1.034232e-09 -9.465974e-08 -6.838703e-11  9.596475e-10  9.884227e-05 -8.657339e-12
   1.713211e-07 -3.135409e-08  6.679834e-06  7.662352e-08 -3.547565e-05  5.372229e-12 -8.123521e-11 -8.657339e-12  4.979570e-05
  -1.054696e-05  3.792651e-05 -3.034532e-05 -2.861504e-05 -8.427618e-04 -8.602667e-07 -1.367857e-08  6.156962e-07 -1.386123e-06
   2.003969e-07 -1.425964e-07 -8.370741e-07  7.600645e-06  1.341364e-06 -4.393243e-12  5.194236e-11 -4.252236e-12  4.671743e-11
  -1.499274e-11  7.854857e-10  2.339292e-09 -9.217518e-09 -1.114732e-07 -3.106779e-11 -4.921463e-12 -4.093247e-11  5.719719e-13
   3.509820e-03  6.288655e-04  1.495065e-02  3.125218e-02  1.127237e+00  1.453673e-07  2.249040e-05 -1.509466e-08  4.368891e-07
  -9.707860e-06 -5.556780e-04 -6.544468e-04 -1.300243e-03 -1.041608e-02 -1.311536e-05  5.012977e-07  2.672512e-07 -9.658215e-08
  -7.300525e-08  5.146078e-07  4.291269e-08 -1.227322e-07 -4.768276e-06  4.820314e-13  5.453332e-09  2.550769e-10 -2.610739e-11
  -5.932053e-06 -1.614580e-04  4.020952e-05  4.633752e-05 -6.933968e-05 -7.668407e-07  8.834278e-06  2.356940e-05 -4.253895e-08
  -7.077476e-04  1.558477e-03  1.444831e-03  1.320263e-03 -4.555168e-01 -3.500834e-08  9.356883e-07 -1.003001e-07  3.971857e-05
  -8.923597e-05  2.756093e-04 -7.109094e-04 -1.102395e-04  4.779052e-03 -3.076337e-07 -6.528338e-08 -4.735573e-07 -8.086701e-07
   2.574902e-03 -4.581377e-03 -2.346068e-02  1.549201e-01  2.718140e-03 -3.362726e-08 -9.790871e-07 -1.622065e-07  1.520009e-06
   3.431112e-05 -5.316147e-05 -1.396779e-04  4.920129e-04 -4.287208e-03 -4.068778e-07 -4.541917e-08  5.998016e-07  3.411355e-08
   9.390624e-03 -6.615195e-03  3.618671e-03  7.266066e-03  1.177138e+00  5.964580e-08 -1.210774e-05 -5.177681e-09  7.990606e-07
   2.434248e-05  6.609656e-04 -3.915952e-04 -7.499049e-04 -3.760280e-03 -1.291120e-05  5.432571e-07  5.372020e-07  1.289701e-07
  -7.062227e-09  5.564182e-07  1.033234e-08 -5.792939e-08 -8.249509e-07  2.345140e-14  1.139855e-09  4.205959e-11 -3.116915e-13
   3.903213e-06 -6.477310e-04 -1.815440e-05 -6.088381e-06 -4.884515e-04  9.593307e-08  1.493096e-06  4.818068e-08 -9.992954e-09
   3.115348e-03  2.132111e-03  1.352156e-01 -1.175406e-03 -5.057107e-01  5.385678e-08  4.783646e-07  3.819546e-08  1.006639e-04
  -4.604464e-05  1.139803e-04 -1.282560e-04 -6.860115e-05 -7.964329e-04 -2.304576e-08  8.222316e-08  1.779671e-06 -1.002472e-06
   1.558984e-03 -1.003144e-03 -5.995985e-03  3.570878e-02  1.423375e-02 -3.193661e-08 -1.638009e-07 -7.021900e-08  1.523949e-07
   5.134865e-06  6.069936e-05  2.647251e-06  7.234074e-05 -5.333258e-04 -2.069499e-07  2.231097e-08 -2.755798e-07 -5.477091e-09
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -1.054696e-05  2.003969e-07 -1.499274e-11  3.509820e-03 -9.707860e-06 -7.300525e-08 -5.932053e-06 -7.077476e-04 -8.923597e-05
   3.792651e-05 -1.425964e-07  7.854857e-10  6.288655e-04 -5.556780e-04  5.146078e-07 -1.614580e-04  1.558477e-03  2.756093e-04
  -3.034532e-05 -8.370741e-07  2.339292e-09  1.495065e-02 -6.544468e-04  4.291269e-08  4.020952e-05  1.444831e-03 -7.109094e-04
  -2.861504e-05  7.600645e-06 -9.217518e-09  3.125218e-02 -1.300243e-03 -1.227322e-07  4.633752e-05  1.320263e-03 -1.102395e-04
  -8.427618e-04  1.341364e-06 -1.114732e-07  1.127237e+00 -1.041608e-02 -4.768276e-06 -6.933968e-05 -4.555168e-01  4.779052e-03
  -8.602667e-07 -4.393243e-12 -3.106779e-11  1.453673e-07 -1.311536e-05  4.820314e-13 -7.668407e-07 -3.500834e-08 -3.076337e-07
  -1.367857e-08  5.194236e-11 -4.921463e-12  2.249040e-05  5.012977e-07  5.453332e-09  8.834278e-06  9.356883e-07 -6.528338e-08
   6.156962e-07 -4.252236e-12 -4.093247e-11 -1.509466e-08  2.672512e-07  2.550769e-10  2.356940e-05 -1.003001e-07 -4.735573e-07
  -1.386123e-06  4.671743e-11  5.719719e-13  4.368891e-07 -9.658215e-08 -2.610739e-11 -4.253895e-08  3.971857e-05 -8.086701e-07
   5.760482e-01  1.437690e-09  5.069660e-07 -2.421134e-04  3.062472e-02  2.906488e-10 -1.953441e-03 -1.307026e-02  1.178416e-01
   1.437690e-09  5.014721e-05 -4.704187e-11  1.030611e-06 -1.009065e-07  2.989235e-11 -2.788878e-08  1.612004e-07 -3.554663e-08
   5.069660e-07 -4.704187e-11  4.804656e-05 -9.409657e-08 -3.151384e-06 -1.566416e-12 -4.682400e-07  1.223735e-08 -6.081620e-07
  -2.421134e-04  1.030611e-06 -9.409657e-08  6.794266e+00 -6.291566e-02  6.258204e-06 -8.587898e-04  3.271116e-01 -2.506704e-03
   3.062472e-02 -1.009065e-07 -3.151384e-06 -6.291566e-02  3.076708e+00 -2.134762e-07 -1.080663e-01  1.051239e-03  5.651103e-01
   2.906488e-10  2.989235e-11 -1.566416e-12  6.258204e-06 -2.134762e-07  8.126588e-05  2.811071e-06 -6.707074e-07 -2.549178e-08
  -1.953441e-03 -2.788878e-08 -4.682400e-07 -8.587898e-04 -1.080663e-01  2.811071e-06  1.944245e+00 -1.078956e-03 -2.320891e-02
  -1.307026e-02  1.612004e-07  1.223735e-08  3.271116e-01  1.051239e-03 -6.707074e-07 -1.078956e-03  4.128406e+00 -2.873579e-02
   1.178416e-01 -3.554663e-08 -6.081620e-07 -2.506704e-03  5.651103e-01 -2.549178e-08 -2.320891e-02 -2.873579e-02  1.728153e+00
   1.263586e-04  8.024275e-05 -1.211157e-06  6.461461e-02  5.772070e-03  1.180867e-06 -3.104182e-03  9.241038e-03  1.399384e-04
   1.718503e-02  3.268509e-07  1.623763e-05  1.177054e-03 -2.948569e-01 -2.358916e-08  7.017461e-04 -1.652604e-03 -2.466149e-02
  -1.121160e-04  1.324575e-06 -2.505665e-08  5.183900e-01  2.299418e-03 -9.767059e-06  5.922714e-03 -6.524077e-02 -2.211722e-03
  -2.384250e-02 -9.721940e-09 -1.089392e-06  1.807326e-04 -5.062277e-01  1.489871e-07 -7.590568e-02 -4.373616e-03 -4.438168e-02
  -1.220245e-09  4.728952e-12 -2.513248e-13  1.761597e-06 -2.740813e-08  5.643390e-10  8.199244e-07 -1.316959e-07  3.095109e-10
   1.731180e-03 -3.290458e-08  1.621072e-07  1.384428e-03  7.006297e-03  6.870456e-07  2.411514e-01 -7.747128e-04 -1.822861e-03
  -2.358396e-02  7.891941e-09  2.561801e-08 -8.523358e-02 -3.331782e-03  5.283543e-07 -1.996321e-04  7.346558e-01 -2.621675e-02
   4.292356e-01  1.005476e-07  9.305561e-07  2.052212e-03  5.520382e-02  6.316423e-08  1.110856e-02 -1.591250e-02  4.165585e-02
   1.783067e-04  1.861471e-05 -6.765457e-08  4.274393e-02 -1.099702e-03  3.508441e-07 -1.096266e-03  1.273165e-02 -1.664953e-04
  -2.481815e-03 -5.075595e-07 -9.644422e-06 -1.381915e-04 -3.439098e-02 -5.187161e-09  9.322316e-03  8.798651e-04 -4.028499e-03
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   2.574902e-03  3.431112e-05  9.390624e-03  2.434248e-05 -7.062227e-09  3.903213e-06  3.115348e-03 -4.604464e-05  1.558984e-03
  -4.581377e-03 -5.316147e-05 -6.615195e-03  6.609656e-04  5.564182e-07 -6.477310e-04  2.132111e-03  1.139803e-04 -1.003144e-03
  -2.346068e-02 -1.396779e-04  3.618671e-03 -3.915952e-04  1.033234e-08 -1.815440e-05  1.352156e-01 -1.282560e-04 -5.995985e-03
   1.549201e-01  4.920129e-04  7.266066e-03 -7.499049e-04 -5.792939e-08 -6.088381e-06 -1.175406e-03 -6.860115e-05  3.570878e-02
   2.718140e-03 -4.287208e-03  1.177138e+00 -3.760280e-03 -8.249509e-07 -4.884515e-04 -5.057107e-01 -7.964329e-04  1.423375e-02
  -3.362726e-08 -4.068778e-07  5.964580e-08 -1.291120e-05  2.345140e-14  9.593307e-08  5.385678e-08 -2.304576e-08 -3.193661e-08
  -9.790871e-07 -4.541917e-08 -1.210774e-05  5.432571e-07  1.139855e-09  1.493096e-06  4.783646e-07  8.222316e-08 -1.638009e-07
  -1.622065e-07  5.998016e-07 -5.177681e-09  5.372020e-07  4.205959e-11  4.818068e-08  3.819546e-08  1.779671e-06 -7.021900e-08
   1.520009e-06  3.411355e-08  7.990606e-07  1.289701e-07 -3.116915e-13 -9.992954e-09  1.006639e-04 -1.002472e-06  1.523949e-07
   1.263586e-04  1.718503e-02 -1.121160e-04 -2.384250e-02 -1.220245e-09  1.731180e-03 -2.358396e-02  4.292356e-01  1.783067e-04
   8.024275e-05  3.268509e-07  1.324575e-06 -9.721940e-09  4.728952e-12 -3.290458e-08  7.891941e-09  1.005476e-07  1.861471e-05
  -1.211157e-06  1.623763e-05 -2.505665e-08 -1.089392e-06 -2.513248e-13  1.621072e-07  2.561801e-08  9.305561e-07 -6.765457e-08
   6.461461e-02  1.177054e-03  5.183900e-01  1.807326e-04  1.761597e-06  1.384428e-03 -8.523358e-02  2.052212e-03  4.274393e-02
   5.772070e-03 -2.948569e-01  2.299418e-03 -5.062277e-01 -2.740813e-08  7.006297e-03 -3.331782e-03  5.520382e-02 -1.099702e-03
   1.180867e-06 -2.358916e-08 -9.767059e-06  1.489871e-07  5.643390e-10  6.870456e-07  5.283543e-07  6.316423e-08  3.508441e-07
  -3.104182e-03  7.017461e-04  5.922714e-03 -7.590568e-02  8.199244e-07  2.411514e-01 -1.996321e-04  1.110856e-02 -1.096266e-03
   9.241038e-03 -1.652604e-03 -6.524077e-02 -4.373616e-03 -1.316959e-07 -7.747128e-04  7.346558e-01 -1.591250e-02  1.273165e-02
   1.399384e-04 -2.466149e-02 -2.211722e-03 -4.438168e-02  3.095109e-10 -1.822861e-03 -2.621675e-02  4.165585e-02 -1.664953e-04
   1.204879e+01  5.075651e-02  2.047897e-03 -5.124266e-04 -9.542007e-08 -1.342538e-03  1.415988e-02  4.306628e-03  6.073223e-01
   5.075651e-02  3.927953e+00  2.938385e-03 -5.482496e-03 -8.474216e-10  8.302022e-03  3.622568e-03  6.036037e-02  5.459726e-04
   2.047897e-03  2.938385e-03  5.253549e+00 -9.446540e-02 -2.770429e-06  4.655369e-03  3.543724e-01 -2.601057e-04  1.486508e-02
  -5.124266e-04 -5.482496e-03 -9.446540e-02  5.228505e+00  4.117630e-08 -2.338422e-02  7.998041e-03 -2.825450e-01  2.825477e-03
  -9.542007e-08 -8.474216e-10 -2.770429e-06  4.117630e-08  4.232964e-05  3.791556e-07  1.429038e-07 -5.143331e-09  2.612542e-08
  -1.342538e-03  8.302022e-03  4.655369e-03 -2.338422e-02  3.791556e-07  3.303829e-01 -2.757107e-05 -1.592641e-02 -6.388198e-04
   1.415988e-02  3.622568e-03  3.543724e-01  7.998041e-03  1.429038e-07 -2.757107e-05  7.047789e+00  3.920881e-02 -2.015045e-02
   4.306628e-03  6.036037e-02 -2.601057e-04 -2.825450e-01 -5.143331e-09 -1.592641e-02  3.920881e-02  4.157009e+00  3.170395e-03
   6.073223e-01  5.459726e-04  1.486508e-02  2.825477e-03  2.612542e-08 -6.388198e-04 -2.015045e-02  3.170395e-03  1.052083e+00
  -1.372430e-02 -1.379502e-01 -3.114564e-03 -1.439501e-01  5.252809e-09  1.533555e-03 -1.316348e-04 -4.642684e-02 -3.354339e-02
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
   5.134865e-06  6.397193e-03 -3.881823e-04 -7.318165e-04 -2.462069e-04  1.101810e-02 -1.270744e-03 -3.402498e+00
   6.069936e-05  6.556140e-02 -1.092968e-03  5.447388e-04 -2.672326e-04  4.562983e-02 -8.177297e-04 -5.465188e-01
   2.647251e-06  2.031523e-02 -1.859782e-03  8.025038e-04 -1.340330e-03 -1.486732e-02 -2.903254e-03 -3.007914e+00
   7.234074e-05 -1.722543e-02  2.073071e-04  4.126243e-04  1.019301e-04 -4.213763e-02  5.020988e-04 -1.406375e+00
  -5.333258e-04  1.660567e-01  1.649776e-03  7.112889e-03  1.032062e-04  1.149102e-01  6.970067e-03 -2.908826e+01
  -2.069499e-07 -8.734538e-08 -4.167430e-07 -5.639110e-09 -1.883950e-06 -9.756557e-08 -1.763993e-06 -1.681312e-05
   2.231097e-08 -2.474486e-06  1.021627e-07  2.333183e-06 -2.563879e-06 -6.837531e-06  2.960311e-07 -6.723084e-04
  -2.755798e-07 -5.942942e-07  2.578692e-06  5.444024e-08 -2.703685e-05 -4.243503e-07  5.526711e-06 -1.497752e-04
  -5.477091e-09 -4.143716e-07  1.482318e-07 -3.058104e-07 -1.557605e-07 -1.681420e-05  1.573426e-07 -4.019202e-04
  -2.481815e-03  1.241285e-03  2.619755e-02  5.854083e-05 -3.492073e-02 -1.735037e-03 -9.700632e-02 -1.658026e+00
  -5.075595e-07 -3.665525e-06 -1.005309e-07 -8.478177e-08  5.576537e-08  8.478240e-06 -2.991636e-07 -3.618042e-04
  -9.644422e-06 -1.880848e-07  8.482525e-07 -1.105431e-08  1.398607e-06 -8.618280e-08 -1.857649e-06 -5.606233e-05
  -1.381915e-04  8.254296e-01  4.601638e-03 -2.692902e-03 -1.434962e-03  3.365676e+00  1.492851e-02 -3.346585e+01
  -3.439098e-02 -4.176978e-03  5.980169e-01 -1.145863e-03  9.740293e-01 -1.359460e-02  5.023445e+00 -1.407590e+01
  -5.187161e-09  1.239965e-06  6.664766e-08  2.802890e-06 -7.163280e-07 -1.134036e-06  4.977103e-08 -4.154677e-04
   9.322316e-03 -3.243691e-03  2.762248e-02 -3.788045e-04  1.269557e-01 -1.260627e-03 -6.142758e-02 -4.671427e+00
   8.798651e-04  4.186766e-01  1.684687e-03  1.657153e-02 -9.174442e-05  7.222583e-01 -7.916400e-03 -2.320523e+01
  -4.028499e-03 -9.943348e-03  9.112610e-02  4.422773e-04 -6.022639e-02 -1.566132e-02  1.051310e+00 -7.523286e+00
  -1.372430e-02 -2.586328e-02 -1.853680e-03 -2.111150e-02 -4.733460e-03  3.344158e-01  1.086730e-02 -3.682817e+01
  -1.379502e-01 -1.376107e-02  6.899110e-02 -5.329938e-04  3.404328e-01 -3.567675e-02 -2.515323e-01 -6.715082e+00
  -3.114564e-03  1.872148e-01 -5.416453e-03  1.451644e-01 -2.333628e-02  2.134095e-01 -2.663004e-02 -1.869814e+01
  -1.439501e-01 -4.117005e-03 -4.954632e-02  1.360726e-03 -4.579150e-03  5.629381e-03  2.531043e-02 -7.720738e+00
   5.252809e-09 -1.823139e-07  1.215848e-08  4.940700e-07 -2.520812e-07 -2.782813e-06  2.810575e-08 -8.234459e-05
   1.533555e-03 -2.622506e-04  5.478100e-02  2.969339e-04  1.417136e-01  1.052615e-03  1.003159e-01 -1.715349e+00
  -1.316348e-04  1.381253e-01  8.494358e-04  8.871161e-03 -2.250868e-02  6.198031e-02 -2.767862e-03 -2.469630e+01
  -4.642684e-02 -2.446288e-03 -1.446948e-02  7.373763e-04  4.650351e-02 -5.211658e-03 -1.255930e-01 -7.074164e+00
  -3.354339e-02 -8.240811e-02 -4.419511e-03  1.417983e-02 -2.291816e-02  6.337225e-02 -1.076768e-02 -7.568868e+00
   2.619163e-01 -1.003260e-03  5.591196e-02 -4.582528e-04  3.074007e-01 -3.332089e-03  2.035330e-01 -1.713075e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  6565 / 6565 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 7.80e+00       1    0.47
s(measurement.no):toneBis.ord1                      9.00e+00 2.09e+00       1    0.51
s(measurement.no):toneBis.ord2                      9.00e+00 7.02e+00       1    0.51
s(measurement.no):toneBis.ord3                      9.00e+00 3.81e+00       1    0.46
s(measurement.no,speaker)                           1.00e+02 5.82e+01       1    0.47
s(measurement.no,speaker):toneBis.ord1              1.00e+02 2.02e-03       1    0.49
s(measurement.no,speaker):toneBis.ord2              1.00e+02 3.32e+00       1    0.51
s(measurement.no,speaker):toneBis.ord3              1.00e+02 1.03e-03       1    0.54
s(measurement.no,speakerLeftRightTone)              5.90e+02 9.51e+01       1    0.49
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.90e+02 9.34e+00       1    0.59
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.90e+02 6.15e+01       1    0.50
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.90e+02 8.71e+01       1    0.52
s(measurement.no,speakerPos)                        3.00e+02 5.28e+01       1    0.54
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 3.43e+00       1    0.54
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 6.35e+01       1    0.51
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 1.86e+01       1    0.52
s(measurement.no,word)                              6.30e+02 1.55e+02       1    0.53
s(measurement.no,wordPos)                           9.99e+02 9.50e+01       1    0.52
s(measurement.no,wordLeftRightTone)                 9.36e+02 4.16e+02       1    0.53

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1a, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F1 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale

# output: 1000 dpi png
# png("pred1-1t.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1a, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(500, 800), xlab = "Time (ms)", ylab = "F1 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1a, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# dev.off()

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1a, view="measurement.no", comp=list(toneBis.ord = c("1","2")), rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 7.474747

plot_diff(gamm.model1a, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.505051 - 6.666667

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1a, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 10.000000

plot_diff(gamm.model1a, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1a, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.505051 - 4.141414
    9.393939 - 10.000000

plot_diff(gamm.model1a, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F1 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1a, pred = list(toneBis.ord=c("1", "2", "3", "4")), main = "Tone", xlab = "F1 (Z)")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.2 Model 1B: /ai/ female, F1 as output

gamm.model1b.noAR <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1b.noAR, paste("Gamm_model1b_noAR.rds"))

gamm.model1b.noAR <- 
  readRDS("Gamm_model1b_noAR.rds")

r.gamm.model1b <- start_value_rho(gamm.model1b.noAR)

# Auto-regressive model
gamm.model1b <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                      
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                    
                      data=data.ai.fem, method="fREML", rho = r.gamm.model1b, AR.start = data.ai.fem$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1b, paste("Gamm_model1b.rds"))

gamm.model1b <- 
  readRDS("Gamm_model1b.rds")

summary(gamm.model1b, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.26984    0.05890   4.581 4.68e-06 ***
toneBis.ord1 -0.48869    0.17880  -2.733  0.00628 ** 
toneBis.ord2  0.04083    0.09876   0.413  0.67933    
toneBis.ord3  0.04743    0.10469   0.453  0.65050    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df      F p-value    
s(measurement.no)              7.797  7.971 89.711 < 2e-16 ***
s(measurement.no):toneBis.ord1 1.001  1.001  0.028 0.86825    
s(measurement.no):toneBis.ord2 7.300  7.841  6.282 < 2e-16 ***
s(measurement.no):toneBis.ord3 4.857  5.606  3.375 0.00424 ** 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.771   Deviance explained = 79.7%
fREML = 6300.4  Scale est. = 0.19144   n = 11198

gam.check(gamm.model1b)

Method: fREML   Optimizer: perf chol
$grad
 [1]  2.355005e-12 -8.137698e-05 -1.851852e-13  3.777201e-12 -4.923351e-11 -4.879585e-05 -4.059321e-05 -1.375344e-12 -1.144684e-11
[10]  4.327649e-13 -1.660183e-11 -8.174095e-05 -2.022915e-10 -9.237056e-14 -4.430824e-09  6.265921e-11  1.762146e-11  8.757439e-13
[19] -1.168843e-12 -4.153122e-12  1.362466e-11  2.014389e-12 -4.909594e-05 -7.184224e-05 -5.618617e-12 -4.156675e-13 -2.515321e-12
[28] -9.318935e-14 -1.172111e-10 -1.527667e-12 -1.002221e-10 -3.694822e-13 -8.341772e-11  2.131628e-12  7.228664e-09

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.601137e+00 -1.658270e-06 -1.500839e-02  1.716137e-02 -1.741485e-01  3.548981e-09 -5.321124e-08  3.473436e-06  1.643944e-03
  -1.658270e-06  8.136709e-05 -1.838063e-07 -1.714651e-06 -1.636336e-05  6.046390e-13  6.116640e-10 -1.011945e-08  1.495573e-06
  -1.500839e-02 -1.838063e-07  2.403274e+00 -5.053821e-02  5.070324e-03  7.838185e-10  2.182950e-07 -2.199431e-05  2.457410e-01
   1.716137e-02 -1.714651e-06 -5.053821e-02  1.327189e+00  3.962985e-02  3.034273e-09 -2.925270e-07  2.361555e-05 -1.432663e-02
  -1.741485e-01 -1.636336e-05  5.070324e-03  3.962985e-02  2.792433e+01 -3.353338e-07  4.282282e-06  4.672041e-04 -1.212472e+00
   3.548981e-09  6.046390e-13  7.838185e-10  3.034273e-09 -3.353338e-07  4.879533e-05  2.739949e-12  2.360583e-07 -4.942857e-08
  -5.321124e-08  6.116640e-10  2.182950e-07 -2.925270e-07  4.282282e-06  2.739949e-12  4.059652e-05 -7.128441e-07  1.022340e-06
   3.473436e-06 -1.011945e-08 -2.199431e-05  2.361555e-05  4.672041e-04  2.360583e-07 -7.128441e-07  5.008349e-01  1.227562e-04
   1.643944e-03  1.495573e-06  2.457410e-01 -1.432663e-02 -1.212472e+00 -4.942857e-08  1.022340e-06  1.227562e-04  6.917649e+00
   9.266282e-05  2.602601e-08  2.938147e-06 -7.657451e-05 -9.745880e-03 -7.342619e-07 -3.708387e-09  5.635391e-03 -6.004610e-02
  -1.245482e-03  1.984711e-06 -1.753096e-02  7.855268e-02 -3.367306e-01  6.080004e-08  1.891240e-06  1.075790e-03 -2.398628e-01
  -3.946488e-09  9.956430e-12  5.023396e-08  1.086675e-07 -2.999885e-08 -1.065017e-10  5.175128e-12  6.647325e-07 -4.387100e-07
  -1.033129e-02  6.355302e-06  2.184217e-02 -9.237799e-03  1.315280e+00 -8.662127e-08 -3.163345e-06  8.143635e-04  9.635374e-02
  -6.065313e-04 -8.329858e-08  7.077922e-05  3.046692e-04 -1.218111e-02 -3.209553e-05  1.880704e-07 -1.875265e-02  1.082522e-03
  -1.499136e-03  1.307646e-05  2.590818e-05 -4.003726e-03  6.238301e-02  3.145407e-08  3.529904e-05 -4.962326e-03  1.221641e-02
   6.947661e-05  1.565759e-08  5.900104e-05  8.612800e-05 -2.437264e-03  4.886513e-08 -9.529832e-07  6.279609e-01 -1.414936e-03
  -2.059964e-03 -5.084830e-07  8.499699e-02 -2.962121e-03 -3.273709e-01 -7.513308e-08  3.511674e-07  1.090234e-03  1.247745e+00
  -1.448993e-04  7.859647e-08 -2.043263e-03 -3.757041e-05 -7.333375e-03 -8.696236e-07 -4.478908e-08 -2.364637e-02 -2.725987e-02
  -1.499063e-03  1.138373e-06  3.838135e-03 -7.183592e-02 -2.630463e-01  5.971007e-08  9.437760e-07  3.745483e-04 -6.939689e-02
   5.211523e-05  4.511044e-08 -2.021804e-04  2.152288e-03  1.353023e-06 -7.000572e-07  1.211288e-07  1.085302e-03 -4.970429e-03
  -2.301350e-02 -1.126469e-05  3.727947e-02  3.243855e-02  1.683618e+00 -2.668195e-08 -2.136182e-06 -1.819414e-03  7.800961e-03
   3.503530e-05  1.079398e-07  6.413040e-04  7.873861e-05  2.473793e-03 -4.365781e-06 -6.354947e-08 -5.576388e-02  1.063948e-04
  -1.247481e-07  2.822664e-10  1.996593e-09 -1.003524e-07  4.967523e-06  9.589899e-13  1.906920e-09 -2.507796e-07  2.899235e-07
   1.996622e-08  3.467387e-11  9.869884e-08  8.543162e-08  4.775805e-07 -8.137000e-13 -2.562308e-10  1.410154e-04  4.417038e-07
   1.076177e-03  1.410646e-06 -1.929480e-03  3.076541e-04 -1.363309e-01  2.090586e-09 -3.721350e-08  4.557519e-05  8.622472e-01
  -7.351581e-05  2.359124e-08 -2.959133e-04 -3.928286e-04 -4.541395e-03 -3.524666e-07 -8.837160e-09 -8.707878e-03 -2.388133e-02
   3.603762e-04  2.021176e-07 -1.234630e-02  5.645724e-02 -6.557462e-02  1.373248e-08  1.094486e-06 -1.760750e-05 -1.054830e-01
   1.862056e-06  4.629065e-10  7.085197e-06  2.393356e-05 -5.865257e-05 -6.675688e-09  1.568708e-09  1.067505e-04 -1.171516e-04
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   9.266282e-05 -1.245482e-03 -3.946488e-09 -1.033129e-02 -6.065313e-04 -1.499136e-03  6.947661e-05 -2.059964e-03 -1.448993e-04
   2.602601e-08  1.984711e-06  9.956430e-12  6.355302e-06 -8.329858e-08  1.307646e-05  1.565759e-08 -5.084830e-07  7.859647e-08
   2.938147e-06 -1.753096e-02  5.023396e-08  2.184217e-02  7.077922e-05  2.590818e-05  5.900104e-05  8.499699e-02 -2.043263e-03
  -7.657451e-05  7.855268e-02  1.086675e-07 -9.237799e-03  3.046692e-04 -4.003726e-03  8.612800e-05 -2.962121e-03 -3.757041e-05
  -9.745880e-03 -3.367306e-01 -2.999885e-08  1.315280e+00 -1.218111e-02  6.238301e-02 -2.437264e-03 -3.273709e-01 -7.333375e-03
  -7.342619e-07  6.080004e-08 -1.065017e-10 -8.662127e-08 -3.209553e-05  3.145407e-08  4.886513e-08 -7.513308e-08 -8.696236e-07
  -3.708387e-09  1.891240e-06  5.175128e-12 -3.163345e-06  1.880704e-07  3.529904e-05 -9.529832e-07  3.511674e-07 -4.478908e-08
   5.635391e-03  1.075790e-03  6.647325e-07  8.143635e-04 -1.875265e-02 -4.962326e-03  6.279609e-01  1.090234e-03 -2.364637e-02
  -6.004610e-02 -2.398628e-01 -4.387100e-07  9.635374e-02  1.082522e-03  1.221641e-02 -1.414936e-03  1.247745e+00 -2.725987e-02
   8.126747e-01 -7.606616e-04 -6.675646e-06  1.379214e-04  9.499431e-02 -2.970089e-04  6.543495e-03 -1.462481e-02  3.983100e-01
  -7.606616e-04  2.283679e+00 -2.873409e-07 -1.081056e-01  8.082037e-03  2.508995e-02  1.768535e-03 -7.271369e-02 -8.990342e-04
  -6.675646e-06 -2.873409e-07  8.175851e-05 -1.262544e-07 -1.508594e-05  2.400389e-07 -2.629687e-06 -4.197506e-07 -1.503728e-05
   1.379214e-04 -1.081056e-01 -1.262544e-07  8.527338e+00 -8.190319e-02  3.099895e-01 -4.756327e-03  9.156217e-01  1.364860e-02
   9.499431e-02  8.082037e-03 -1.508594e-05 -8.190319e-02  6.810330e+00 -6.597013e-05 -1.737594e-01 -1.576030e-03  1.701012e+00
  -2.970089e-04  2.508995e-02  2.400389e-07  3.099895e-01 -6.597013e-05  1.711563e+00 -7.959894e-02 -3.956474e-02 -1.094120e-03
   6.543495e-03  1.768535e-03 -2.629687e-06 -4.756327e-03 -1.737594e-01 -7.959894e-02  3.001638e+00  2.267123e-03 -7.995519e-03
  -1.462481e-02 -7.271369e-02 -4.197506e-07  9.156217e-01 -1.576030e-03 -3.956474e-02  2.267123e-03  8.978132e+00 -1.249457e-01
   3.983100e-01 -8.990342e-04 -1.503728e-05  1.364860e-02  1.701012e+00 -1.094120e-03 -7.995519e-03 -1.249457e-01  5.481277e+00
  -5.167945e-04  1.058850e+00 -6.593845e-07 -2.320689e-01  3.083562e-03 -2.078027e-03 -1.177973e-04  1.202266e-01 -4.466178e-03
  -2.516821e-03  1.315711e-02  1.572683e-04  2.300866e-03 -3.890745e-01  5.075512e-03 -4.160621e-02 -3.156635e-03 -3.805164e-02
   3.228192e-03  9.070257e-02  4.847512e-07  5.198314e-01 -1.976906e-02  3.254096e-03 -5.663102e-03  1.563297e-01  1.833771e-03
  -2.204157e-02 -1.770631e-03  1.006857e-06 -1.535049e-02 -3.201078e-01 -1.873451e-03 -3.014200e-02 -2.601196e-03  1.484434e-01
  -9.378718e-09  9.070615e-07  4.952687e-12  4.974743e-06 -8.814002e-09  4.812767e-05 -1.516241e-06 -5.506462e-07 -7.372308e-08
  -5.549307e-08  3.236098e-07  3.137854e-10  4.611403e-07 -3.508587e-05 -6.267245e-06  2.206592e-04  4.656689e-07  5.278224e-07
  -9.744508e-03 -3.928118e-02 -1.773240e-07  9.325692e-02  4.847993e-03  4.869862e-05  2.720506e-04  1.107136e+00 -2.082365e-03
   2.941717e-01  1.252069e-04 -1.045957e-06 -2.191842e-04 -4.032050e-02  7.991828e-04 -3.501802e-02 -1.646201e-02 -3.063455e-02
  -3.146631e-04  1.013963e+00  5.164392e-07  4.364294e-02  7.812063e-04  5.944861e-03  3.862983e-04 -3.819058e-02 -1.699113e-03
  -4.770183e-04 -2.733978e-04  2.080351e-06 -3.624303e-05  4.232058e-03  7.185098e-05 -1.942619e-04 -8.952064e-05 -1.754698e-03
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
  -1.499063e-03  5.211523e-05 -2.301350e-02  3.503530e-05 -1.247481e-07  1.996622e-08  1.076177e-03 -7.351581e-05  3.603762e-04
   1.138373e-06  4.511044e-08 -1.126469e-05  1.079398e-07  2.822664e-10  3.467387e-11  1.410646e-06  2.359124e-08  2.021176e-07
   3.838135e-03 -2.021804e-04  3.727947e-02  6.413040e-04  1.996593e-09  9.869884e-08 -1.929480e-03 -2.959133e-04 -1.234630e-02
  -7.183592e-02  2.152288e-03  3.243855e-02  7.873861e-05 -1.003524e-07  8.543162e-08  3.076541e-04 -3.928286e-04  5.645724e-02
  -2.630463e-01  1.353023e-06  1.683618e+00  2.473793e-03  4.967523e-06  4.775805e-07 -1.363309e-01 -4.541395e-03 -6.557462e-02
   5.971007e-08 -7.000572e-07 -2.668195e-08 -4.365781e-06  9.589899e-13 -8.137000e-13  2.090586e-09 -3.524666e-07  1.373248e-08
   9.437760e-07  1.211288e-07 -2.136182e-06 -6.354947e-08  1.906920e-09 -2.562308e-10 -3.721350e-08 -8.837160e-09  1.094486e-06
   3.745483e-04  1.085302e-03 -1.819414e-03 -5.576388e-02 -2.507796e-07  1.410154e-04  4.557519e-05 -8.707878e-03 -1.760750e-05
  -6.939689e-02 -4.970429e-03  7.800961e-03  1.063948e-04  2.899235e-07  4.417038e-07  8.622472e-01 -2.388133e-02 -1.054830e-01
  -5.167945e-04 -2.516821e-03  3.228192e-03 -2.204157e-02 -9.378718e-09 -5.549307e-08 -9.744508e-03  2.941717e-01 -3.146631e-04
   1.058850e+00  1.315711e-02  9.070257e-02 -1.770631e-03  9.070615e-07  3.236098e-07 -3.928118e-02  1.252069e-04  1.013963e+00
  -6.593845e-07  1.572683e-04  4.847512e-07  1.006857e-06  4.952687e-12  3.137854e-10 -1.773240e-07 -1.045957e-06  5.164392e-07
  -2.320689e-01  2.300866e-03  5.198314e-01 -1.535049e-02  4.974743e-06  4.611403e-07  9.325692e-02 -2.191842e-04  4.364294e-02
   3.083562e-03 -3.890745e-01 -1.976906e-02 -3.201078e-01 -8.814002e-09 -3.508587e-05  4.847993e-03 -4.032050e-02  7.812063e-04
  -2.078027e-03  5.075512e-03  3.254096e-03 -1.873451e-03  4.812767e-05 -6.267245e-06  4.869862e-05  7.991828e-04  5.944861e-03
  -1.177973e-04 -4.160621e-02 -5.663102e-03 -3.014200e-02 -1.516241e-06  2.206592e-04  2.720506e-04 -3.501802e-02  3.862983e-04
   1.202266e-01 -3.156635e-03  1.563297e-01 -2.601196e-03 -5.506462e-07  4.656689e-07  1.107136e+00 -1.646201e-02 -3.819058e-02
  -4.466178e-03 -3.805164e-02  1.833771e-03  1.484434e-01 -7.372308e-08  5.278224e-07 -2.082365e-03 -3.063455e-02 -1.699113e-03
   3.050479e+00  8.247993e-02 -2.082275e-02  1.275849e-03 -5.567415e-07  2.037004e-07 -4.574614e-03  1.482445e-03  3.809748e-01
   8.247993e-02  6.647249e+00 -2.112107e-03  3.854187e-02  4.188705e-08  1.618703e-05 -2.676884e-04  7.951103e-04 -7.995715e-03
  -2.082275e-02 -2.112107e-03  1.304513e+01  1.916612e-02  4.560937e-06  5.387971e-06  3.823348e-01 -1.507341e-03  4.830968e-01
   1.275849e-03  3.854187e-02  1.916612e-02  4.109909e+00  4.683115e-07 -5.331128e-05 -1.366595e-04 -3.264444e-01 -1.391543e-03
  -5.567415e-07  4.188705e-08  4.560937e-06  4.683115e-07  4.909899e-05 -6.918195e-10 -4.324559e-07  2.538614e-08 -4.871085e-08
   2.037004e-07  1.618703e-05  5.387971e-06 -5.331128e-05 -6.918195e-10  7.201659e-05  4.082391e-07 -1.800783e-05 -5.573377e-07
  -4.574614e-03 -2.676884e-04  3.823348e-01 -1.366595e-04 -4.324559e-07  4.082391e-07  1.542742e+00 -7.014322e-02 -7.428371e-02
   1.482445e-03  7.951103e-04 -1.507341e-03 -3.264444e-01  2.538614e-08 -1.800783e-05 -7.014322e-02  1.942995e+00  4.301125e-04
   3.809748e-01 -7.995715e-03  4.830968e-01 -1.391543e-03 -4.871085e-08 -5.573377e-07 -7.428371e-02  4.301125e-04  1.932947e+00
   6.335997e-05  1.730895e-02  2.780325e-04  3.992384e-03  2.957308e-09  1.889700e-07 -2.844252e-05 -3.138603e-04  1.387570e-03
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
   1.862056e-06  1.224010e-02 -1.534033e-03 -1.034786e-03 -2.553204e-04  1.583492e-02 -5.814198e-04 -3.398287e+00
   4.629065e-10 -1.408013e-05  4.747407e-07 -3.288173e-06  2.103919e-07 -6.468130e-06  2.184187e-07 -1.770537e-04
   7.085197e-06 -4.184055e-03 -4.935098e-03 -4.154366e-03 -5.556190e-04 -7.116186e-03 -2.383510e-04 -3.149811e+00
   2.393356e-05 -2.630989e-02 -2.349461e-03 -9.347061e-03 -1.662483e-03  2.355889e-02 -6.160756e-04 -1.928526e+00
  -5.865257e-05  1.106368e-01 -2.181461e-02  3.241771e-02 -2.806085e-03 -3.091672e-02 -1.624398e-03 -3.279171e+01
  -6.675688e-09 -3.967256e-08  1.452411e-06  4.751684e-08 -9.941705e-08 -9.451054e-08 -2.870830e-09 -1.216870e-05
   1.568708e-09 -7.555576e-06  4.712337e-08  3.644105e-06 -8.158573e-08 -1.148077e-06 -3.064838e-08 -2.522826e-04
   1.067505e-04 -7.463498e-04  3.216701e-02  1.180157e-03 -1.304944e-02 -1.377268e-03 -6.056157e-02 -1.383545e+00
  -1.171516e-04  3.327595e-01 -9.691961e-03  6.433185e-02  1.126673e-03  3.686700e-01 -1.606973e-02 -1.566177e+01
  -4.770183e-04  6.047906e-03 -2.688093e-02  1.954468e-03 -1.923574e-03  6.194074e-03  7.054727e-02 -1.959279e+00
  -2.733978e-04 -8.459636e-02  1.046164e-02  1.196766e-01 -1.404408e-02 -6.377852e-02  6.315453e-03 -8.971754e+00
   2.080351e-06  4.401494e-07 -5.520421e-05 -7.427578e-07  2.978219e-05  8.652885e-07 -6.137451e-06 -2.826292e-04
  -3.624303e-05  9.323632e-01  4.698917e-03  5.625916e-01  4.619305e-03  3.012858e+00 -6.138344e-03 -3.764725e+01
   4.232058e-03  4.373885e-02  3.747390e+00  1.127466e-02  1.083238e+00  4.967543e-02  5.568526e+00 -2.271341e+01
   7.185098e-05  1.289479e-01  3.037234e-03  1.433014e-01  3.544926e-04  6.865017e-02 -1.403029e-03 -8.711404e+00
  -1.942619e-04 -3.758715e-04 -4.129769e-02 -9.751264e-04 -9.994131e-02  5.227153e-03  2.487336e-02 -5.281231e+00
  -8.952064e-05  5.665497e-01 -5.547123e-03  3.485268e-02  2.457342e-03  1.102789e+00 -1.143414e-04 -3.549148e+01
  -1.754698e-03  8.422340e-04 -5.197211e-01 -6.649523e-03 -5.139187e-01 -3.612642e-02 -4.957946e-01 -1.203109e+01
   6.335997e-05  1.111576e-01  9.646000e-03  3.378472e-01 -5.855717e-03 -2.432590e-02  3.665522e-03 -1.662720e+01
   1.730895e-02  9.250286e-03 -6.560844e-01 -1.824690e-04  9.562812e-03 -1.831421e-02 -6.251386e-01 -9.175892e+00
   2.780325e-04  9.945617e-02  9.242423e-03  3.256043e-02 -1.061381e-02  3.016209e-01 -9.741670e-03 -3.164751e+01
   3.992384e-03  2.645248e-03  1.127626e-01  5.492350e-03 -5.203557e-01 -4.624418e-03  2.827055e-02 -6.367651e+00
   2.957308e-09  6.934881e-06  9.372916e-08  1.130848e-05 -1.215490e-07  4.211097e-06 -7.804410e-08 -3.477411e-04
   1.889700e-07 -1.149069e-06  2.028897e-05  3.436174e-07  3.525361e-06 -6.453213e-06  1.122015e-04 -1.092900e-03
  -2.844252e-05  1.292142e-01  2.840407e-03  4.025624e-02 -3.553123e-03  2.960472e-01 -1.347774e-02 -1.136728e+01
  -3.138603e-04  4.838612e-03  1.321379e-02 -5.882335e-03  1.735786e-02  1.155861e-02 -9.853477e-02 -5.040649e+00
   1.387570e-03  2.223193e-01  1.340519e-03  1.920210e-01  1.640177e-02  2.777977e-01 -6.187345e-03 -1.184315e+01
   2.078162e-03 -3.044769e-05 -1.032302e-03  7.814940e-04  1.162611e-02  3.874255e-04  8.491506e-03 -1.115762e-01
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  6427 / 6427 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 7.80e+00       1    0.64
s(measurement.no):toneBis.ord1                      9.00e+00 1.00e+00       1    0.55
s(measurement.no):toneBis.ord2                      9.00e+00 7.30e+00       1    0.52
s(measurement.no):toneBis.ord3                      9.00e+00 4.86e+00       1    0.52
s(measurement.no,speaker)                           1.00e+02 6.56e+01       1    0.66
s(measurement.no,speaker):toneBis.ord1              1.00e+02 2.77e+00       1    0.56
s(measurement.no,speaker):toneBis.ord2              1.00e+02 3.52e+01       1    0.55
s(measurement.no,speaker):toneBis.ord3              1.00e+02 1.79e+01       1    0.51
s(measurement.no,speakerLeftRightTone)              5.90e+02 1.21e+02       1    0.59
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.90e+02 2.80e+01       1    0.59
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.90e+02 9.50e+01       1    0.55
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.90e+02 5.16e+01       1    0.56
s(measurement.no,speakerPos)                        3.00e+02 7.60e+01       1    0.60
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 3.12e-03       1    0.51
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 3.28e+01       1    0.65
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 2.39e+01       1    0.59
s(measurement.no,word)                              6.15e+02 2.59e+02       1    0.56
s(measurement.no,wordPos)                           9.33e+02 1.71e+02       1    0.56
s(measurement.no,wordLeftRightTone)                 8.79e+02 2.80e+02       1    0.55

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1b, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F1 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1b, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1b, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1b, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1b, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(500, 1000), xlab = "Time (ms)", ylab = "F1 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2f * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1b, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT1f * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1b, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT4f * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1b, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3f * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1b, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 7.575758

#axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1b, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    1.010101 - 9.494949

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1b, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 10.000000

#axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1b, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.707071 - 1.111111

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1b, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.505051 - 2.222222

#axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1b, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F1 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1b, pred = list(toneBis.ord=c("1", "2", "3", "4")), main = "Tone", xlab = "F1 (Z)")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.3 Model 1C: /au/ male, F1 as output

gamm.model1c.noAR <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1c.noAR, paste("Gamm_model1c_noAR.rds"))

gamm.model1c.noAR <- 
  readRDS("Gamm_model1c_noAR.rds")

r.gamm.model1c <- start_value_rho(gamm.model1c.noAR)

# Auto-regressive model

gamm.model1c <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.mas, method="fREML", rho = r.gamm.model1c, AR.start = data.au.mas$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1c, paste("Gamm_model1c.rds"))

gamm.model1c <- 
  readRDS("Gamm_model1c.rds")

summary(gamm.model1c, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.060211   0.055331   1.088 0.276537    
toneBis.ord1 -0.442326   0.091890  -4.814 1.51e-06 ***
toneBis.ord2  0.006794   0.091474   0.074 0.940793    
toneBis.ord3  0.312667   0.081333   3.844 0.000122 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df      F p-value    
s(measurement.no)              7.670  8.032 64.474  <2e-16 ***
s(measurement.no):toneBis.ord1 3.091  3.802  2.286  0.0513 .  
s(measurement.no):toneBis.ord2 1.000  1.000  2.899  0.0887 .  
s(measurement.no):toneBis.ord3 3.533  4.096  2.729  0.0266 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.657   Deviance explained = 70.9%
fREML =  10543  Scale est. = 0.32426   n = 10857

gam.check(gamm.model1c)

Method: fREML   Optimizer: perf chol
$grad
 [1]  9.681145e-14 -1.776357e-14 -5.014119e-05 -1.232348e-13  0.000000e+00 -4.773879e-05 -9.651113e-05 -4.919213e-05 -7.603781e-05
[10] -9.746363e-05 -9.508623e-05 -9.405792e-05 -3.765876e-13 -2.202682e-13  3.046452e-13 -1.048051e-13  1.376677e-13 -9.797667e-05
[19]  1.705303e-13 -2.309264e-14  6.750156e-14 -5.329071e-14  4.618528e-14  4.973799e-14  7.105427e-15 -4.810617e-05  1.456613e-13
[28]  1.127987e-13  7.105427e-14  1.037392e-12  3.410605e-13  4.263256e-13  3.694822e-13 -1.989520e-13  4.547474e-12

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.462193e+00 -1.137325e-02  2.190058e-06  4.961225e-02 -3.299671e-02  3.034639e-10 -1.601364e-06 -1.059766e-07 -1.905974e-07
  -1.137325e-02  7.165497e-01  8.770145e-06 -8.565477e-03 -3.409280e-02  3.928110e-09  4.405821e-06  2.557979e-07  6.709955e-07
   2.190058e-06  8.770145e-06  5.015151e-05 -1.162515e-06  6.877865e-06 -8.786020e-12 -1.388045e-10  3.015069e-11  3.691683e-10
   4.961225e-02 -8.565477e-03 -1.162515e-06  5.975867e-01 -1.085729e-02 -1.070045e-08  9.489704e-07 -2.866613e-08  6.375498e-07
  -3.299671e-02 -3.409280e-02  6.877865e-06 -1.085729e-02  1.136755e+01 -1.040605e-07  1.266081e-05 -4.021332e-06  1.595450e-05
   3.034639e-10  3.928110e-09 -8.786020e-12 -1.070045e-08 -1.040605e-07  4.774418e-05 -9.201469e-12 -2.021643e-08 -4.675430e-12
  -1.601364e-06  4.405821e-06 -1.388045e-10  9.489704e-07  1.266081e-05 -9.201469e-12  9.652650e-05  2.090843e-09 -1.368250e-10
  -1.059766e-07  2.557979e-07  3.015069e-11 -2.866613e-08 -4.021332e-06 -2.021643e-08  2.090843e-09  6.114258e-05 -2.836031e-10
  -1.905974e-07  6.709955e-07  3.691683e-10  6.375498e-07  1.595450e-05 -4.675430e-12 -1.368250e-10 -2.836031e-10  7.604259e-05
  -6.461417e-09 -3.339083e-09 -3.010039e-11  7.812958e-09  4.948893e-08 -2.882737e-10 -2.544260e-12 -5.151366e-09 -2.826756e-11
   1.658825e-06  2.316011e-07  1.950866e-10  1.054927e-05 -3.005954e-05 -4.152255e-12  2.043134e-10  2.765375e-11  6.381021e-11
  -1.185780e-08  8.986651e-08 -2.662848e-11 -1.803007e-07  2.231221e-06 -1.277243e-08 -5.635538e-11  2.693756e-08  1.804706e-11
  -3.902424e-02  3.928589e-02 -1.283390e-05  1.435327e-03  2.311166e+00 -3.138625e-07  2.727351e-06 -1.025585e-05  5.034014e-06
  -1.084292e-03  7.067769e-05  9.671237e-09 -6.501510e-04 -1.721403e-03  5.485718e-05 -1.151871e-07 -6.720659e-04  2.549189e-07
  -2.301347e-03  2.187482e-02 -4.970189e-08  2.126738e-03  1.201095e-02  9.626930e-09  2.978077e-05  1.308316e-06 -1.200650e-07
  -8.358042e-05  1.642326e-04  1.237021e-08 -9.268128e-06  1.168928e-03 -4.752375e-06  5.296669e-07  2.028409e-03  1.479344e-07
  -6.800272e-03  1.899594e-03 -4.565243e-06  1.394357e-02  1.095966e-01 -3.464066e-08 -1.244823e-07 -1.655263e-06  3.196528e-05
  -2.752269e-11  2.697647e-09 -2.787166e-12  4.287188e-09  4.705936e-08 -1.709127e-10 -6.774337e-13  1.289242e-10 -2.199913e-12
   4.770773e-02  1.085158e-02 -2.860897e-07  1.282609e-01 -6.229080e-01 -3.956074e-08  1.627322e-06 -7.930093e-07  2.448632e-06
  -1.977190e-04  2.066826e-04 -3.862087e-08 -7.248829e-04  3.348742e-03 -9.762976e-06 -5.757055e-08  2.384311e-04 -4.479960e-07
  -1.343996e-02 -1.933901e-03  2.930288e-06  7.188260e-03  1.661145e+00 -3.937988e-08  7.578434e-06  2.148657e-06  4.615502e-06
  -4.175626e-05 -2.101576e-04 -8.880872e-09  2.622473e-05  7.348834e-04  2.890917e-05 -1.242873e-07  2.482589e-05 -2.224812e-07
  -9.111643e-03  6.695449e-02 -1.554034e-06  5.871734e-03  3.604937e-02 -6.445434e-08  1.026974e-04  4.622120e-06 -1.045842e-06
  -6.853272e-05  2.821799e-04 -1.037149e-08  9.065814e-07 -3.353190e-04 -4.377071e-06  2.009601e-07  4.006969e-03 -1.215765e-08
  -7.935187e-03  1.563530e-02 -2.224423e-06  2.513018e-02  3.115543e-01  1.865849e-07 -4.243435e-06 -6.480917e-06  1.123289e-04
   2.606433e-10  1.204878e-09  6.592714e-13  1.075133e-09 -1.999768e-09 -4.865325e-11  7.665352e-14 -1.229003e-09 -4.183841e-12
   6.499681e-03 -4.337589e-03  2.201209e-06  1.410892e-02 -6.364897e-02 -3.788359e-08  1.781038e-06 -5.354775e-07  1.069530e-06
  -1.745212e-05  1.488390e-04 -1.441488e-07  1.126533e-04  1.484850e-03 -3.247099e-06 -5.826576e-08  6.727364e-06  1.530749e-07
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -6.461417e-09  1.658825e-06 -1.185780e-08 -3.902424e-02 -1.084292e-03 -2.301347e-03 -8.358042e-05 -6.800272e-03 -2.752269e-11
  -3.339083e-09  2.316011e-07  8.986651e-08  3.928589e-02  7.067769e-05  2.187482e-02  1.642326e-04  1.899594e-03  2.697647e-09
  -3.010039e-11  1.950866e-10 -2.662848e-11 -1.283390e-05  9.671237e-09 -4.970189e-08  1.237021e-08 -4.565243e-06 -2.787166e-12
   7.812958e-09  1.054927e-05 -1.803007e-07  1.435327e-03 -6.501510e-04  2.126738e-03 -9.268128e-06  1.394357e-02  4.287188e-09
   4.948893e-08 -3.005954e-05  2.231221e-06  2.311166e+00 -1.721403e-03  1.201095e-02  1.168928e-03  1.095966e-01  4.705936e-08
  -2.882737e-10 -4.152255e-12 -1.277243e-08 -3.138625e-07  5.485718e-05  9.626930e-09 -4.752375e-06 -3.464066e-08 -1.709127e-10
  -2.544260e-12  2.043134e-10 -5.635538e-11  2.727351e-06 -1.151871e-07  2.978077e-05  5.296669e-07 -1.244823e-07 -6.774337e-13
  -5.151366e-09  2.765375e-11  2.693756e-08 -1.025585e-05 -6.720659e-04  1.308316e-06  2.028409e-03 -1.655263e-06  1.289242e-10
  -2.826756e-11  6.381021e-11  1.804706e-11  5.034014e-06  2.549189e-07 -1.200650e-07  1.479344e-07  3.196528e-05 -2.199913e-12
   9.746247e-05  4.434396e-13  1.665822e-09 -8.896769e-09 -5.043191e-06 -1.296528e-08  1.014107e-06 -5.554128e-08 -5.180392e-10
   4.434396e-13  9.509938e-05 -3.973733e-10  2.038503e-06 -1.961299e-07  1.149578e-07  4.186410e-08  1.077729e-06  6.584054e-13
   1.665822e-09 -3.973733e-10  9.524929e-05  3.431335e-06 -1.703278e-04 -2.878891e-08  1.575130e-05 -1.290375e-06  1.675771e-10
  -8.896769e-09  2.038503e-06  3.431335e-06  2.001702e+01  1.743489e-02 -4.870241e-02 -1.206872e-03  5.231005e-02  2.583869e-08
  -5.043191e-06 -1.961299e-07 -1.703278e-04  1.743489e-02  9.366168e+00  5.405320e-04 -6.586277e-01 -1.088053e-03 -6.508741e-06
  -1.296528e-08  1.149578e-07 -2.878891e-08 -4.870241e-02  5.405320e-04  6.505597e-01 -1.409563e-03  5.937920e-03 -1.546145e-09
   1.014107e-06  4.186410e-08  1.575130e-05 -1.206872e-03 -6.586277e-01 -1.409563e-03  4.192710e+00  6.599362e-04  2.120674e-06
  -5.554128e-08  1.077729e-06 -1.290375e-06  5.231005e-02 -1.088053e-03  5.937920e-03  6.599362e-04  8.007949e-01  2.794845e-07
  -5.180392e-10  6.584054e-13  1.675771e-10  2.583869e-08 -6.508741e-06 -1.546145e-09  2.120674e-06  2.794845e-07  9.797595e-05
   2.185791e-07  1.977680e-04 -1.017408e-05  6.595814e-01 -7.083191e-03 -3.142783e-02  1.157587e-03  1.022589e-01  1.679006e-07
   3.398476e-07 -5.695386e-07  1.129604e-03 -3.413994e-03  3.122517e-01 -1.108743e-04  1.268524e-02 -5.110197e-03 -3.336018e-06
  -4.308363e-08 -1.531934e-06 -7.510140e-08  1.163879e+00  2.002509e-03  7.062957e-02  1.162798e-03  4.974677e-02 -5.655830e-08
   3.752938e-06 -6.565319e-08 -8.537799e-05 -7.451213e-03  2.989293e-01  3.930083e-04  9.161272e-02 -2.647663e-04  1.672484e-06
  -4.062828e-08  1.311150e-07 -2.898343e-07  3.627772e-01 -2.117975e-03  4.955032e-01  3.697365e-03 -8.676228e-04 -2.151922e-08
  -2.975014e-06  1.571743e-08  2.503284e-05 -2.175810e-03 -1.405529e-01  1.248606e-03  9.639966e-01 -2.523207e-04 -1.551930e-06
   3.419304e-07  2.935498e-06  8.654007e-08  8.892519e-02  7.448568e-03  1.048710e-03  2.689088e-03  1.339327e+00  3.882685e-07
  -4.160150e-10  2.721495e-13  3.098299e-10  3.447925e-08 -2.874961e-06  1.095669e-09  2.781406e-07  9.935389e-08 -3.532816e-10
   3.881944e-08  4.176235e-05 -3.425326e-07 -9.770851e-02 -2.060665e-04 -9.758117e-03  6.341571e-04  1.611548e-02 -1.690358e-08
   2.737885e-06 -9.798108e-08  3.428581e-04  1.557291e-03  4.989887e-02 -1.606503e-04  5.026723e-02  1.216896e-03  1.071718e-06
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   4.770773e-02 -1.977190e-04 -1.343996e-02 -4.175626e-05 -9.111643e-03 -6.853272e-05 -7.935187e-03  2.606433e-10  6.499681e-03
   1.085158e-02  2.066826e-04 -1.933901e-03 -2.101576e-04  6.695449e-02  2.821799e-04  1.563530e-02  1.204878e-09 -4.337589e-03
  -2.860897e-07 -3.862087e-08  2.930288e-06 -8.880872e-09 -1.554034e-06 -1.037149e-08 -2.224423e-06  6.592714e-13  2.201209e-06
   1.282609e-01 -7.248829e-04  7.188260e-03  2.622473e-05  5.871734e-03  9.065814e-07  2.513018e-02  1.075133e-09  1.410892e-02
  -6.229080e-01  3.348742e-03  1.661145e+00  7.348834e-04  3.604937e-02 -3.353190e-04  3.115543e-01 -1.999768e-09 -6.364897e-02
  -3.956074e-08 -9.762976e-06 -3.937988e-08  2.890917e-05 -6.445434e-08 -4.377071e-06  1.865849e-07 -4.865325e-11 -3.788359e-08
   1.627322e-06 -5.757055e-08  7.578434e-06 -1.242873e-07  1.026974e-04  2.009601e-07 -4.243435e-06  7.665352e-14  1.781038e-06
  -7.930093e-07  2.384311e-04  2.148657e-06  2.482589e-05  4.622120e-06  4.006969e-03 -6.480917e-06 -1.229003e-09 -5.354775e-07
   2.448632e-06 -4.479960e-07  4.615502e-06 -2.224812e-07 -1.045842e-06 -1.215765e-08  1.123289e-04 -4.183841e-12  1.069530e-06
   2.185791e-07  3.398476e-07 -4.308363e-08  3.752938e-06 -4.062828e-08 -2.975014e-06  3.419304e-07 -4.160150e-10  3.881944e-08
   1.977680e-04 -5.695386e-07 -1.531934e-06 -6.565319e-08  1.311150e-07  1.571743e-08  2.935498e-06  2.721495e-13  4.176235e-05
  -1.017408e-05  1.129604e-03 -7.510140e-08 -8.537799e-05 -2.898343e-07  2.503284e-05  8.654007e-08  3.098299e-10 -3.425326e-07
   6.595814e-01 -3.413994e-03  1.163879e+00 -7.451213e-03  3.627772e-01 -2.175810e-03  8.892519e-02  3.447925e-08 -9.770851e-02
  -7.083191e-03  3.122517e-01  2.002509e-03  2.989293e-01 -2.117975e-03 -1.405529e-01  7.448568e-03 -2.874961e-06 -2.060665e-04
  -3.142783e-02 -1.108743e-04  7.062957e-02  3.930083e-04  4.955032e-01  1.248606e-03  1.048710e-03  1.095669e-09 -9.758117e-03
   1.157587e-03  1.268524e-02  1.162798e-03  9.161272e-02  3.697365e-03  9.639966e-01  2.689088e-03  2.781406e-07  6.341571e-04
   1.022589e-01 -5.110197e-03  4.974677e-02 -2.647663e-04 -8.676228e-04 -2.523207e-04  1.339327e+00  9.935389e-08  1.611548e-02
   1.679006e-07 -3.336018e-06 -5.655830e-08  1.672484e-06 -2.151922e-08 -1.551930e-06  3.882685e-07 -3.532816e-10 -1.690358e-08
   2.964601e+01 -1.000070e-01  3.990149e-02 -5.343230e-04  4.815307e-02 -1.788876e-03  3.013415e-01  4.373562e-08  1.243523e+00
  -1.000070e-01  6.707622e+00  1.852905e-03 -1.718423e-01 -8.109850e-04  1.798271e-01 -1.339314e-02  8.008443e-08 -1.617912e-03
   3.990149e-02  1.852905e-03  5.918355e+00  1.942569e-03  3.829888e-01  1.545561e-03  7.380914e-03 -1.491648e-08  1.600251e-02
  -5.343230e-04 -1.718423e-01  1.942569e-03  6.570015e+00  1.604866e-03 -4.767542e-02  1.001515e-02 -1.523000e-06 -1.859622e-03
   4.815307e-02 -8.109850e-04  3.829888e-01  1.604866e-03  3.092613e+00 -1.530331e-02 -2.120323e-02 -9.188110e-10 -2.002501e-02
  -1.788876e-03  1.798271e-01  1.545561e-03 -4.767542e-02 -1.530331e-02  4.062483e+00 -3.936288e-03 -2.616709e-07 -4.062657e-07
   3.013415e-01 -1.339314e-02  7.380914e-03  1.001515e-02 -2.120323e-02 -3.936288e-03  1.052413e+01  7.457004e-08  5.812132e-02
   4.373562e-08  8.008443e-08 -1.491648e-08 -1.523000e-06 -9.188110e-10 -2.616709e-07  7.457004e-08  4.810591e-05 -1.994496e-08
   1.243523e+00 -1.617912e-03  1.600251e-02 -1.859622e-03 -2.002501e-02 -4.062657e-07  5.812132e-02 -1.994496e-08  2.421344e+00
  -3.335378e-03  8.666964e-01  1.443970e-03  3.468851e-02 -1.473467e-03  1.383896e-01  7.312562e-03 -2.579624e-08 -9.417563e-03
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -1.745212e-05 -1.669524e-02 -6.842249e-05 -7.692942e-03  3.536344e-05 -7.395802e-03 -1.103745e-03 -3.334809e+00
   1.488390e-04  4.834621e-02  2.880422e-04  6.522554e-03  2.700369e-05  3.033396e-02  1.087881e-03 -1.045698e+00
  -1.441488e-07 -7.018557e-06 -1.224287e-07 -4.522978e-06  6.144557e-08  1.476941e-05  5.821859e-08 -8.303867e-05
   1.126533e-04  5.311030e-03 -1.395670e-04 -1.229322e-02  5.245256e-05 -2.802305e-02 -7.220352e-04 -1.266666e+00
   1.484850e-03 -3.492194e-02 -1.918000e-04 -9.375976e-02 -1.558337e-03 -5.439919e-02 -3.417869e-04 -1.923601e+01
  -3.247099e-06  2.241967e-10 -2.302079e-06  4.555519e-08 -9.834699e-06 -2.230898e-07 -2.020533e-05 -1.622524e-04
  -5.826576e-08  1.978920e-05 -9.588048e-09  1.279239e-05 -3.429814e-08  4.291544e-05  3.837361e-08 -6.159935e-04
   6.727364e-06  6.768032e-06 -2.782608e-04  7.837338e-06  2.265946e-04  6.880751e-06 -1.052839e-03 -7.417941e-03
   1.530749e-07 -2.616508e-06  1.061744e-07  6.628547e-06 -5.070010e-08  4.828187e-07  1.804501e-07 -4.386036e-04
   2.737885e-06 -2.672274e-08  1.133633e-06 -1.171547e-07 -1.085175e-05 -1.680063e-07 -2.984472e-05 -7.142378e-05
  -9.798108e-08 -1.711964e-06 -8.437386e-08  4.018060e-06  6.170792e-08 -8.677162e-06 -5.626829e-08 -6.636234e-04
   3.428581e-04  8.893011e-07  5.857168e-05  2.865924e-06  5.802056e-04  1.071035e-06 -2.763050e-04 -2.605385e-03
   1.557291e-03  1.586671e+00 -3.409431e-04  1.224802e+00 -1.578324e-03  7.357650e+00 -8.204265e-03 -5.987854e+01
   4.989887e-02 -5.537058e-03  1.463651e-01 -6.000875e-03 -1.844886e-01 -2.148919e-02  4.863955e+00 -2.140589e+01
  -1.606503e-04  6.908516e-02 -4.340261e-04  9.699562e-02 -1.316896e-03  2.688983e-01 -2.653502e-03 -6.857947e+00
   5.026723e-02 -3.398816e-03 -4.013640e-02 -1.580185e-03 -1.308629e-02 -7.266771e-03 -1.339855e-01 -8.612105e+00
   1.216896e-03  2.046038e-01 -1.286860e-04  2.672029e-01 -4.576968e-04  3.382203e-01 -2.216111e-03 -6.115421e+00
   1.071718e-06 -2.787125e-08 -3.605416e-06 -4.352495e-08 -2.179201e-05 -3.075142e-08 -2.106535e-05 -6.610294e-05
  -3.335378e-03  1.165285e+00  2.881074e-05  2.344594e+00  2.675007e-02  1.345475e+00  9.786980e-03 -7.893145e+01
   8.666964e-01  3.878495e-03 -5.850941e-03 -4.293601e-03  1.477834e+00  1.449560e-02 -1.886093e-01 -1.404010e+01
   1.443970e-03 -3.627841e-02 -1.818683e-05 -1.630241e-01 -2.907367e-03 -3.827110e-01 -3.012385e-03 -2.082925e+01
   3.468851e-02  6.373469e-04 -1.267363e-01  1.523656e-04 -8.546051e-01 -3.760326e-03 -8.728003e-01 -9.027570e+00
  -1.473467e-03  1.934407e-01 -4.226946e-05  2.078162e-01  1.718289e-03  6.697884e-01 -1.164509e-03 -1.339107e+01
   1.383896e-01  5.182333e-04 -4.928362e-02 -1.290755e-03  1.153150e-01  1.040919e-03 -9.274989e-01 -6.860890e+00
   7.312562e-03  3.106654e-01 -2.499850e-03  4.053729e-01 -1.459706e-02  1.086589e+00 -2.373529e-02 -2.832217e+01
  -2.579624e-08 -1.967350e-09 -1.210182e-06 -4.161774e-09 -8.641286e-06 -6.208313e-08 -1.117334e-05 -3.064790e-05
  -9.417563e-03  3.523924e-02 -8.187868e-04 -2.319515e-01 -5.228767e-03  1.896648e-01  2.048313e-03 -1.435757e+01
   1.815132e+00  1.200168e-03  6.309525e-02  3.454065e-03 -9.499130e-02 -6.578861e-03  5.878848e-01 -5.296680e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  7802 / 7802 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 7.67e+00       1    0.41
s(measurement.no):toneBis.ord1                      9.00e+00 3.09e+00       1    0.36
s(measurement.no):toneBis.ord2                      9.00e+00 1.00e+00       1    0.43
s(measurement.no):toneBis.ord3                      9.00e+00 3.53e+00       1    0.45
s(measurement.no,speaker)                           1.00e+02 3.85e+01       1    0.38
s(measurement.no,speaker):toneBis.ord1              1.00e+02 1.64e-02       1    0.44
s(measurement.no,speaker):toneBis.ord2              1.00e+02 1.37e-03       1    0.48
s(measurement.no,speaker):toneBis.ord3              1.00e+02 6.92e-03       1    0.39
s(measurement.no,speakerLeftRightTone)              6.00e+02 1.63e+02       1    0.41
s(measurement.no,speakerLeftRightTone):toneBis.ord1 6.00e+02 3.09e+01       1    0.43
s(measurement.no,speakerLeftRightTone):toneBis.ord2 6.00e+02 1.22e+01       1    0.43
s(measurement.no,speakerLeftRightTone):toneBis.ord3 6.00e+02 1.86e+02       1    0.48
s(measurement.no,speakerPos)                        3.00e+02 5.97e+01       1    0.46
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 4.05e+01       1    0.40
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 5.66e+01       1    0.43
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 3.93e+01       1    0.38
s(measurement.no,word)                              9.75e+02 1.52e+02       1    0.46
s(measurement.no,wordPos)                           1.43e+03 3.31e+02       1    0.42
s(measurement.no,wordLeftRightTone)                 1.36e+03 5.31e+02       1    0.50

#Plotting
# Normalized scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1c, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F1 (Z)", xaxt = "n", font.lab = 2, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1c, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1c, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1c, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1c, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(500, 700), xlab = "Time (ms)", ylab = "F1 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1au * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1c, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2au * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1c, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3au * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1c, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4au * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1c, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 8.686869

plot_diff(gamm.model1c, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.303030 - 10.000000

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1c, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.404040 - 10.000000

plot_diff(gamm.model1c, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    3.535354 - 10.000000

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1c, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1c, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    1.515152 - 10.000000

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F1 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1c, pred = list(toneBis.ord=c("1", "2", "3", "4")), main = "Tone", xlab = "F1 (Z)")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.4 Model 1D: /au/ female, F1 as output

gamm.model1d.noAR <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1d.noAR, paste("Gamm_model1d_noAR.rds"))

gamm.model1d.noAR <- 
  readRDS("Gamm_model1d_noAR.rds")

r.gamm.model1d <- start_value_rho(gamm.model1d.noAR)

# Auto-regressive model

gamm.model1d <- bam(f1Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                    
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.fem, method="fREML", rho = r.gamm.model1d, AR.start = data.au.fem$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1d, paste("Gamm_model1d.rds"))

gamm.model1d <- readRDS("Gamm_model1d.rds")

summary(gamm.model1d, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.16219    0.05724   2.834  0.00461 ** 
toneBis.ord1 -0.35416    0.08495  -4.169 3.09e-05 ***
toneBis.ord2 -0.17573    0.08834  -1.989  0.04671 *  
toneBis.ord3  0.06142    0.07753   0.792  0.42827    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df       F p-value    
s(measurement.no)              8.099  8.313 111.901 < 2e-16 ***
s(measurement.no):toneBis.ord1 4.538  5.432   3.894 0.00152 ** 
s(measurement.no):toneBis.ord2 3.468  4.286   3.950 0.00273 ** 
s(measurement.no):toneBis.ord3 4.246  5.069   2.037 0.07229 .  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.745   Deviance explained = 78.3%
fREML = 8460.7  Scale est. = 0.24131   n = 10252

gam.check(gamm.model1d)

Method: fREML   Optimizer: perf chol
$grad
 [1] -2.263432e-07 -1.976717e-07 -4.406521e-07  1.689388e-07  2.551381e-06 -5.279386e-05  8.773527e-07 -7.287982e-05 -5.097203e-05
[10] -7.403346e-05  6.809239e-07 -6.501874e-05 -9.847376e-06  4.267025e-08 -1.835476e-07  2.827569e-08 -9.938958e-06 -6.767732e-05
[19]  9.200066e-07 -4.559484e-08 -2.151888e-06  6.358038e-08 -1.777578e-07  3.117976e-08 -3.710818e-05 -5.833773e-07 -4.337288e-07
[28]  8.411823e-08 -6.577230e-06 -2.895434e-08 -1.117279e-05  7.061571e-08 -3.450988e-06 -4.460145e-08  2.880854e-04

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.719882e+00 -6.374096e-02  2.505414e-02  4.814090e-02 -9.631971e-02 -1.404865e-09 -9.373219e-03  7.087722e-10  6.127760e-06
  -6.374096e-02  1.001090e+00  3.017618e-02  6.985698e-02  8.837656e-03 -1.631016e-09  1.072949e-01 -6.796644e-09  7.241832e-06
   2.505414e-02  3.017618e-02  7.325820e-01 -3.950766e-02 -4.589143e-02 -6.800776e-10 -4.114420e-04  2.781061e-09  3.820560e-05
   4.814090e-02  6.985698e-02 -3.950766e-02  7.254346e-01  7.524961e-03  2.660483e-09  1.330269e-02 -1.903899e-09 -4.622223e-06
  -9.631971e-02  8.837656e-03 -4.589143e-02  7.524961e-03  1.799996e+01 -1.061054e-07  5.166475e-02 -1.132994e-07 -3.929350e-05
  -1.404865e-09 -1.631016e-09 -6.800776e-10  2.660483e-09 -1.061054e-07  5.279333e-05 -2.978768e-08 -8.590446e-11  9.614664e-12
  -9.373219e-03  1.072949e-01 -4.114420e-04  1.330269e-02  5.166475e-02 -2.978768e-08  1.812253e+00 -1.346346e-06 -3.033064e-05
   7.087722e-10 -6.796644e-09  2.781061e-09 -1.903899e-09 -1.132994e-07 -8.590446e-11 -1.346346e-06  7.288428e-05  1.626612e-11
   6.127760e-06  7.241832e-06  3.820560e-05 -4.622223e-06 -3.929350e-05  9.614664e-12 -3.033064e-05  1.626612e-11  5.135041e-05
   5.314193e-10 -9.257959e-10  2.879306e-08  5.552890e-09 -5.525486e-08 -8.369255e-11  2.143556e-08 -1.518963e-11  5.834618e-10
   2.274547e-03  1.072763e-02  2.492812e-03  1.394692e-01 -1.988880e-01  1.916568e-08  4.792329e-02 -3.243844e-08  9.997226e-06
  -1.111173e-09  5.897139e-10 -8.087530e-09  6.682677e-09 -2.392742e-07 -2.892540e-10  1.091549e-08 -3.465795e-10 -2.007142e-11
   2.620110e-02 -2.996804e-02 -4.894108e-03  3.055260e-02  2.013612e+00  6.695699e-08 -1.561834e-01  2.447529e-08  1.162779e-04
  -7.433868e-04 -6.750249e-04  8.054930e-04 -2.011274e-04 -1.837843e-03 -1.982588e-05 -4.529597e-03  6.760651e-06  5.967540e-08
  -1.248212e-03 -6.480229e-03 -4.324118e-05  1.239154e-03  2.181420e-02 -1.381198e-08  2.625696e-01 -2.007985e-07 -7.033481e-06
  -6.213140e-05 -4.387455e-05  2.414587e-05 -1.674730e-05 -2.742644e-03  1.894041e-06 -2.257622e-02  3.113919e-05  3.807825e-08
   1.267149e-03  2.323124e-05 -7.329733e-03 -1.533247e-03 -4.484158e-02  5.020462e-09  8.008529e-04 -2.855147e-09  4.083725e-05
   4.857004e-10 -1.242279e-09  4.036157e-08 -3.248280e-09  4.243023e-08 -4.978105e-11 -2.582413e-08  1.729028e-10  3.749629e-10
   2.772011e-03  7.646099e-03  4.450527e-03  7.558802e-02 -1.870510e-02  1.731108e-08  1.813999e-02 -2.792952e-08 -8.344105e-06
  -3.504490e-05 -3.015488e-05  4.414684e-04  2.536289e-05  5.772426e-06 -9.965276e-07 -9.064633e-04 -8.995492e-07  1.154059e-06
  -3.839964e-02  1.684916e-02  1.814193e-02  4.255589e-03  1.478484e+00 -9.265753e-08 -8.667577e-02 -3.991575e-08 -2.529648e-06
  -1.765703e-04 -8.811565e-05  7.446677e-04 -1.723752e-04 -2.138417e-03 -1.170593e-05 -4.021861e-04 -4.545869e-06 -1.800377e-06
  -1.731150e-03  3.933000e-02 -7.205234e-04  2.775706e-03  4.568702e-02 -6.025779e-09  4.641221e-01 -2.727564e-07 -1.647334e-05
   1.286741e-05 -1.463523e-04  5.967653e-05 -2.262872e-06  2.740310e-04 -4.034316e-07 -9.891064e-03  4.191823e-05 -3.476107e-07
   2.264014e-07  1.977239e-07  4.407712e-07 -1.689810e-07 -2.552102e-06  1.736332e-13 -8.775873e-07  7.163725e-13  1.315749e-08
   1.750026e-05 -9.473915e-06  7.756894e-04  2.008073e-04  6.817781e-04 -1.007362e-06 -7.565601e-05  1.618769e-06  4.606361e-06
   1.574863e-03  3.081301e-03 -4.526982e-04  2.530293e-02 -5.200849e-02  1.500412e-10  9.609124e-03 -1.002926e-08 -2.884997e-06
  -7.483400e-06  4.348445e-05 -1.223848e-04  1.977409e-04 -1.580909e-03 -1.858509e-06 -3.897562e-04 -1.014167e-06  5.155660e-07
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   5.314193e-10  2.274547e-03 -1.111173e-09  2.620110e-02 -7.433868e-04 -1.248212e-03 -6.213140e-05  1.267149e-03  4.857004e-10
  -9.257959e-10  1.072763e-02  5.897139e-10 -2.996804e-02 -6.750249e-04 -6.480229e-03 -4.387455e-05  2.323124e-05 -1.242279e-09
   2.879306e-08  2.492812e-03 -8.087530e-09 -4.894108e-03  8.054930e-04 -4.324118e-05  2.414587e-05 -7.329733e-03  4.036157e-08
   5.552890e-09  1.394692e-01  6.682677e-09  3.055260e-02 -2.011274e-04  1.239154e-03 -1.674730e-05 -1.533247e-03 -3.248280e-09
  -5.525486e-08 -1.988880e-01 -2.392742e-07  2.013612e+00 -1.837843e-03  2.181420e-02 -2.742644e-03 -4.484158e-02  4.243023e-08
  -8.369255e-11  1.916568e-08 -2.892540e-10  6.695699e-08 -1.982588e-05 -1.381198e-08  1.894041e-06  5.020462e-09 -4.978105e-11
   2.143556e-08  4.792329e-02  1.091549e-08 -1.561834e-01 -4.529597e-03  2.625696e-01 -2.257622e-02  8.008529e-04 -2.582413e-08
  -1.518963e-11 -3.243844e-08 -3.465795e-10  2.447529e-08  6.760651e-06 -2.007985e-07  3.113919e-05 -2.855147e-09  1.729028e-10
   5.834618e-10  9.997226e-06 -2.007142e-11  1.162779e-04  5.967540e-08 -7.033481e-06  3.807825e-08  4.083725e-05  3.749629e-10
   7.403751e-05 -1.185690e-08 -3.068717e-10  4.826966e-08  8.830783e-06 -9.541884e-09  1.130280e-05  1.957801e-07  7.583647e-11
  -1.185690e-08  2.135689e+00 -2.420654e-07  5.300989e-02  1.048772e-03 -3.625253e-04 -1.022922e-03 -1.140626e-02 -2.321139e-08
  -3.068717e-10 -2.420654e-07  6.503692e-05 -1.660778e-08 -1.339911e-05 -4.858821e-09 -1.861709e-06 -3.211166e-08 -4.519375e-10
   4.826966e-08  5.300989e-02 -1.660778e-08  1.516154e+01 -7.181350e-03  2.059961e-01 -6.485742e-05  2.045591e-02 -2.628778e-08
   8.830783e-06  1.048772e-03 -1.339911e-05 -7.181350e-03  1.239728e+01 -1.401804e-03 -6.271258e-02 -2.017129e-03 -9.008563e-06
  -9.541884e-09 -3.625253e-04 -4.858821e-09  2.059961e-01 -1.401804e-03  3.348027e-01 -1.574907e-02  1.516012e-02 -4.376707e-09
   1.130280e-05 -1.022922e-03 -1.861709e-06 -6.485742e-05 -6.271258e-02 -1.574907e-02  8.692207e+00 -1.475071e-03  1.533469e-05
   1.957801e-07 -1.140626e-02 -3.211166e-08  2.045591e-02 -2.017129e-03  1.516012e-02 -1.475071e-03  6.327256e-01  8.234094e-08
   7.583647e-11 -2.321139e-08 -4.519375e-10 -2.628778e-08 -9.008563e-06 -4.376707e-09  1.533469e-05  8.234094e-08  6.767965e-05
  -1.054305e-08  8.104957e-01  1.824848e-07  3.082562e-01  3.074863e-03  3.464327e-04  3.023401e-05  3.049563e-03 -5.680324e-08
   6.048768e-06 -2.728781e-03  7.251506e-05  6.604647e-04 -2.493408e-01 -1.103673e-04 -1.810824e-01  1.044739e-03  3.327389e-07
  -2.301120e-08  6.215264e-02 -1.453757e-07  1.510867e+00  4.552609e-04 -5.803870e-02 -1.371764e-03 -1.040800e-01 -6.639354e-09
  -8.193931e-06  7.033470e-04 -8.228853e-06  4.470595e-03 -5.669122e-01 -6.944822e-05 -1.817126e-01 -1.233467e-03 -5.135050e-06
  -4.023665e-09  1.100889e-02  1.804001e-08  4.340836e-02 -1.353649e-03  1.172621e-01 -6.919525e-03  9.396136e-03 -1.780136e-08
   3.207942e-07 -1.957709e-04 -1.783666e-06 -2.705693e-03  6.245040e-03 -1.742444e-03  4.578193e-01 -3.077189e-04  2.510749e-06
   2.282393e-11 -6.811420e-07 -4.371616e-14  9.850672e-06 -4.268059e-08  1.836178e-07 -2.828188e-08  9.942378e-06  2.054004e-11
   2.838289e-05  2.100485e-03  6.311043e-06  2.275647e-04  1.377342e-01 -2.892137e-05  1.241461e-01  2.727595e-04  2.859461e-05
  -9.703763e-09  4.422858e-01 -5.698338e-08  4.916871e-02  4.320569e-04 -6.348153e-03 -5.537533e-04 -2.798626e-04  3.312658e-08
   3.033710e-06 -6.493487e-04  1.214390e-04 -1.308908e-03 -1.161469e-01  4.329828e-04 -1.067085e-01 -5.521842e-05  4.597474e-06
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   2.772011e-03 -3.504490e-05 -3.839964e-02 -1.765703e-04 -1.731150e-03  1.286741e-05  2.264014e-07  1.750026e-05  1.574863e-03
   7.646099e-03 -3.015488e-05  1.684916e-02 -8.811565e-05  3.933000e-02 -1.463523e-04  1.977239e-07 -9.473915e-06  3.081301e-03
   4.450527e-03  4.414684e-04  1.814193e-02  7.446677e-04 -7.205234e-04  5.967653e-05  4.407712e-07  7.756894e-04 -4.526982e-04
   7.558802e-02  2.536289e-05  4.255589e-03 -1.723752e-04  2.775706e-03 -2.262872e-06 -1.689810e-07  2.008073e-04  2.530293e-02
  -1.870510e-02  5.772426e-06  1.478484e+00 -2.138417e-03  4.568702e-02  2.740310e-04 -2.552102e-06  6.817781e-04 -5.200849e-02
   1.731108e-08 -9.965276e-07 -9.265753e-08 -1.170593e-05 -6.025779e-09 -4.034316e-07  1.736332e-13 -1.007362e-06  1.500412e-10
   1.813999e-02 -9.064633e-04 -8.667577e-02 -4.021861e-04  4.641221e-01 -9.891064e-03 -8.775873e-07 -7.565601e-05  9.609124e-03
  -2.792952e-08 -8.995492e-07 -3.991575e-08 -4.545869e-06 -2.727564e-07  4.191823e-05  7.163725e-13  1.618769e-06 -1.002926e-08
  -8.344105e-06  1.154059e-06 -2.529648e-06 -1.800377e-06 -1.647334e-05 -3.476107e-07  1.315749e-08  4.606361e-06 -2.884997e-06
  -1.054305e-08  6.048768e-06 -2.301120e-08 -8.193931e-06 -4.023665e-09  3.207942e-07  2.282393e-11  2.838289e-05 -9.703763e-09
   8.104957e-01 -2.728781e-03  6.215264e-02  7.033470e-04  1.100889e-02 -1.957709e-04 -6.811420e-07  2.100485e-03  4.422858e-01
   1.824848e-07  7.251506e-05 -1.453757e-07 -8.228853e-06  1.804001e-08 -1.783666e-06 -4.371616e-14  6.311043e-06 -5.698338e-08
   3.082562e-01  6.604647e-04  1.510867e+00  4.470595e-03  4.340836e-02 -2.705693e-03  9.850672e-06  2.275647e-04  4.916871e-02
   3.074863e-03 -2.493408e-01  4.552609e-04 -5.669122e-01 -1.353649e-03  6.245040e-03 -4.268059e-08  1.377342e-01  4.320569e-04
   3.464327e-04 -1.103673e-04 -5.803870e-02 -6.944822e-05  1.172621e-01 -1.742444e-03  1.836178e-07 -2.892137e-05 -6.348153e-03
   3.023401e-05 -1.810824e-01 -1.371764e-03 -1.817126e-01 -6.919525e-03  4.578193e-01 -2.828188e-08  1.241461e-01 -5.537533e-04
   3.049563e-03  1.044739e-03 -1.040800e-01 -1.233467e-03  9.396136e-03 -3.077189e-04  9.942378e-06  2.727595e-04 -2.798626e-04
  -5.680324e-08  3.327389e-07 -6.639354e-09 -5.135050e-06 -1.780136e-08  2.510749e-06  2.054004e-11  2.859461e-05  3.312658e-08
   1.521835e+00 -1.929429e-02  2.078064e-01  1.809842e-03  1.806581e-02 -5.484151e-05 -9.202683e-07  8.076768e-04  3.478483e-01
  -1.929429e-02  3.533783e+00 -1.414744e-04 -7.177971e-02 -4.116702e-05 -4.998375e-02  4.560778e-08  3.067298e-02 -9.397574e-04
   2.078064e-01 -1.414744e-04  1.100136e+01 -5.728799e-02 -6.189906e-02 -1.717053e-03  2.152876e-06  1.411978e-03  1.084166e-01
   1.809842e-03 -7.177971e-02 -5.728799e-02  7.768225e+00 -5.678028e-04 -2.906126e-01 -6.359645e-08 -3.283686e-01  7.331257e-05
   1.806581e-02 -4.116702e-05 -6.189906e-02 -5.678028e-04  4.814646e-01 -1.432874e-02  1.778260e-07 -1.545298e-04 -9.357957e-04
  -5.484151e-05 -4.998375e-02 -1.717053e-03 -2.906126e-01 -1.432874e-02  2.179102e+00 -3.118803e-08 -2.027721e-02 -3.007584e-04
  -9.202683e-07  4.560778e-08  2.152876e-06 -6.359645e-08  1.778260e-07 -3.118803e-08  3.710933e-05  5.835496e-07  4.338878e-07
   8.076768e-04  3.067298e-02  1.411978e-03 -3.283686e-01 -1.545298e-04 -2.027721e-02  5.835496e-07  1.863205e+00  1.048861e-03
   3.478483e-01 -9.397574e-04  1.084166e-01  7.331257e-05 -9.357957e-04 -3.007584e-04  4.338878e-07  1.048861e-03  4.504621e-01
   2.843962e-03  6.433080e-01  5.181299e-04 -7.124836e-01 -5.553771e-04 -1.205477e-02 -8.414271e-08 -9.675473e-02  2.608824e-03
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -7.483400e-06 -6.457061e-03 -4.866757e-05 -2.623354e-03 -9.260853e-05 -4.831239e-03 -1.753270e-04 -3.549456e+00
   4.348445e-05 -8.345307e-03  4.715595e-05  8.782287e-03  6.515888e-05 -4.040887e-03  5.732897e-05 -1.768847e+00
  -1.223848e-04 -2.384022e-02  4.330284e-05 -1.763753e-02  3.073177e-04 -3.411146e-02  7.329211e-05 -1.233854e+00
   1.977409e-04  2.138683e-02  3.516615e-04  7.736704e-03  2.507909e-04  2.353880e-02  6.983143e-04 -1.623143e+00
  -1.580909e-03 -1.299940e-01 -2.276029e-03  5.395768e-02 -4.506029e-03 -7.342013e-02 -3.951828e-03 -2.529763e+01
  -1.858509e-06 -5.310641e-10 -1.771554e-07 -4.051232e-08 -3.798999e-06  1.367845e-08 -2.022776e-06 -2.307532e-05
  -3.897562e-04 -2.469285e-02  9.501564e-04 -8.375349e-02  4.935205e-05  1.500951e-01  2.751987e-03 -7.787842e+00
  -1.014167e-06  3.762137e-08  3.236177e-06 -4.960624e-08 -4.109764e-06  1.126400e-07 -1.702672e-05 -1.641659e-04
   5.155660e-07  2.527288e-04  2.957060e-07  2.029024e-04  1.120360e-06  5.041652e-04  5.001739e-07 -3.366850e-03
   3.033710e-06  2.226380e-07 -1.415280e-05  3.086092e-07 -1.911637e-05  6.383422e-07 -1.106117e-05 -1.185934e-04
  -6.493487e-04  1.679894e-01  1.351288e-03  1.179851e-02  2.084263e-03  1.927557e-01  3.106268e-03 -7.442323e+00
   1.214390e-04  3.980803e-09 -9.799087e-06 -1.120932e-07 -1.869553e-05 -2.645890e-08 -2.936238e-05 -2.848236e-04
  -1.308908e-03  1.874124e+00  2.203628e-03  1.446923e+00  5.731671e-03  5.403847e+00  6.559770e-03 -5.168309e+01
  -1.161469e-01 -3.672947e-03  1.305758e+00 -3.228798e-03  2.411451e+00  7.205885e-03  3.829019e+00 -2.238870e+01
   4.329828e-04  7.811130e-02 -1.716155e-05  1.267694e-01 -1.305139e-03  1.845445e-01 -1.050663e-03 -5.853152e+00
  -1.067085e-01  1.559257e-04 -1.803393e-02 -6.338187e-03  3.902090e-01 -1.368173e-02  2.191919e-01 -1.515289e+01
  -5.521842e-05  1.256109e-01  1.137014e-03  2.296203e-01  1.185579e-03  4.889840e-01  1.460659e-03 -6.578798e+00
   4.597474e-06 -3.991942e-08 -2.941514e-06 -6.237097e-08 -1.745894e-05 -1.349976e-07  5.957818e-06 -1.583714e-04
   2.843962e-03  1.582936e-01  6.043350e-04  1.923116e-01  2.027460e-03  3.824882e-01  3.699380e-04 -1.475162e+01
   6.433080e-01  1.790135e-03  1.624946e-01 -1.844016e-03  1.867315e-02  4.297133e-03  5.963202e-01 -9.144651e+00
   5.181299e-04  2.529623e-01 -1.954315e-03  2.273077e-01 -1.054465e-02  1.289267e+00 -5.948466e-03 -2.824699e+01
  -7.124836e-01 -1.118320e-03 -7.321193e-02 -3.539665e-03 -2.010042e-01 -1.280070e-03 -5.106159e-01 -9.528931e+00
  -5.553771e-04  3.471603e-02 -1.138120e-04 -6.272792e-02 -1.206359e-03  1.103864e-01 -8.097978e-04 -5.086359e+00
  -1.205477e-02 -2.672290e-03 -7.624514e-02 -1.477304e-03 -2.253861e-01 -2.954305e-03 -9.091869e-02 -4.991273e+00
  -8.414271e-08  6.579429e-06  2.896277e-08  1.117641e-05 -7.063274e-08  3.452499e-06  4.461275e-08 -2.881613e-04
  -9.675473e-02  1.275920e-03  4.701303e-02 -6.415411e-03  5.138981e-01  5.410833e-03  3.255360e-01 -4.452017e+00
   2.608824e-03  8.457272e-02  5.461874e-04  8.411937e-02  1.044321e-03  5.224196e-02  8.346469e-04 -6.073009e+00
   5.570895e+00 -5.487112e-04 -1.418358e-01 -6.669694e-03 -2.634647e-01  4.374331e-04 -6.774332e-01 -8.751659e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  7542 / 7542 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 8.10e+00    1.01    0.84
s(measurement.no):toneBis.ord1                      9.00e+00 4.54e+00    1.01    0.82
s(measurement.no):toneBis.ord2                      9.00e+00 3.47e+00    1.01    0.80
s(measurement.no):toneBis.ord3                      9.00e+00 4.25e+00    1.01    0.84
s(measurement.no,speaker)                           1.00e+02 5.06e+01    1.01    0.83
s(measurement.no,speaker):toneBis.ord1              1.00e+02 1.56e+01    1.01    0.83
s(measurement.no,speaker):toneBis.ord2              1.00e+02 7.22e-03    1.01    0.86
s(measurement.no,speaker):toneBis.ord3              1.00e+02 1.49e+01    1.01    0.85
s(measurement.no,speakerLeftRightTone)              5.80e+02 1.48e+02    1.01    0.85
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.80e+02 4.20e+01    1.01    0.83
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.80e+02 1.32e+01    1.01    0.89
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.80e+02 4.78e+01    1.01    0.85
s(measurement.no,speakerPos)                        3.00e+02 7.56e+01    1.01    0.82
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 2.02e+01    1.01    0.86
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 8.90e+00    1.01    0.85
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 2.96e+01    1.01    0.81
s(measurement.no,word)                              9.30e+02 1.84e+02    1.01    0.84
s(measurement.no,wordPos)                           1.39e+03 3.87e+02    1.01    0.89
s(measurement.no,wordLeftRightTone)                 1.27e+03 4.42e+02    1.01    0.86

# Plotting
# Normalized scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1d, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F1 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1d, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1d, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1d, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1d, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(550, 900), xlab = "Time (ms)", ylab = "F1 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2auf * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1d, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1auf * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1d, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3auf * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1d, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4auf * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plots between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1d, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.606061 - 5.959596
    9.191919 - 10.000000

plot_diff(gamm.model1d, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.303030 - 8.888889

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1d, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 8.181818

plot_diff(gamm.model1d, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    6.161616 - 10.000000

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1d, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    6.666667 - 10.000000

plot_diff(gamm.model1d, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F1 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1d, pred = list(toneBis.ord=c("1", "2", "3", "4")), xlab = "F1 (Z)", main = "Tone")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.5 For Appendix B: Organizing tones as “1 < 2 < 3 < 4”

gamm.initialOrder.noAR <- bam(f1Zscore2 ~ tone.ord + 
                        
                        # smooth                        
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=tone.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +  
                    
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.initialOrder.noAR, paste("Gamm_initialOrder_noAR.rds"))

gamm.initialOrder.noAR <- 
  readRDS("Gamm_initialOrder_noAR.rds")

r.gamm.initialOrder <- start_value_rho(gamm.initialOrder.noAR)

# Auto-regressive model

gamm.InitialOrder <- bam(f1Zscore2 ~ tone.ord + 
                       # smooth                        
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=tone.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=tone.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                    
                      data=data.au.mas, method="fREML", rho = r.gamm.initialOrder, AR.start = data.au.mas$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.InitialOrder, paste("Gamm_InitialOrder.rds"))

gamm.InitialOrder <- readRDS("Gamm_f1_toneO_au_masbis2.rds")

summary(gamm.InitialOrder, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ tone.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = tone.ord, bs = "cr") + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = tone.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = tone.ord) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = tone.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) -0.42981    0.07043  -6.103 1.08e-09 ***
tone.ord2    0.49433    0.10315   4.792 1.68e-06 ***
tone.ord3    0.79634    0.10038   7.934 2.38e-15 ***
tone.ord4    0.49171    0.08219   5.983 2.27e-09 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                              edf Ref.df      F p-value    
s(measurement.no)           7.629  8.003 40.127 < 2e-16 ***
s(measurement.no):tone.ord2 1.942  2.239  0.992 0.31439    
s(measurement.no):tone.ord3 3.698  4.288  4.171 0.00183 ** 
s(measurement.no):tone.ord4 2.745  3.322  2.009 0.09147 .  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.655   Deviance explained = 70.7%
fREML =  10549  Scale est. = 0.32538   n = 10857

#Plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.InitialOrder, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(tone.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F1 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * tone.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):tone.ord2,s(measurement.no,speaker):tone.ord3,s(measurement.no,speaker):tone.ord4,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):tone.ord2,s(measurement.no,speakerLeftRightTone):tone.ord3,s(measurement.no,speakerLeftRightTone):tone.ord4,s(measurement.no,speakerPos),s(measurement.no,speakerPos):tone.ord2,s(measurement.no,speakerPos):tone.ord3,s(measurement.no,speakerPos):tone.ord4,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.InitialOrder, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(tone.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * tone.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):tone.ord2,s(measurement.no,speaker):tone.ord3,s(measurement.no,speaker):tone.ord4,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):tone.ord2,s(measurement.no,speakerLeftRightTone):tone.ord3,s(measurement.no,speakerLeftRightTone):tone.ord4,s(measurement.no,speakerPos),s(measurement.no,speakerPos):tone.ord2,s(measurement.no,speakerPos):tone.ord3,s(measurement.no,speakerPos):tone.ord4,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.InitialOrder, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(tone.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * tone.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):tone.ord2,s(measurement.no,speaker):tone.ord3,s(measurement.no,speaker):tone.ord4,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):tone.ord2,s(measurement.no,speakerLeftRightTone):tone.ord3,s(measurement.no,speakerLeftRightTone):tone.ord4,s(measurement.no,speakerPos),s(measurement.no,speakerPos):tone.ord2,s(measurement.no,speakerPos):tone.ord3,s(measurement.no,speakerPos):tone.ord4,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.InitialOrder, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(tone.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * tone.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):tone.ord2,s(measurement.no,speaker):tone.ord3,s(measurement.no,speaker):tone.ord4,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):tone.ord2,s(measurement.no,speakerLeftRightTone):tone.ord3,s(measurement.no,speakerLeftRightTone):tone.ord4,s(measurement.no,speakerPos),s(measurement.no,speakerPos):tone.ord2,s(measurement.no,speakerPos):tone.ord3,s(measurement.no,speakerPos):tone.ord4,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

4.6 Model 1E: /ai/ male, F2 as output

gamm.model1e.noAR <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1e.noAR, paste("Gamm_model1e_noAR.rds"))

gamm.model1e.noAR <- 
  readRDS("Gamm_model1e_noAR.rds")

r.gamm.model1e <- start_value_rho(gamm.model1e.noAR)

# Auto-regressive model

gamm.model1e <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                    
                      data=data.ai.mas, method="fREML", rho = r.gamm.model1e, AR.start = data.ai.mas$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1e, paste("Gamm_model1e.rds"))

gamm.model1e <- readRDS("Gamm_model1e.rds")

summary(gamm.model1e, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
              Estimate Std. Error t value Pr(>|t|)
(Intercept)    0.01529    0.07512   0.204    0.839
toneBis.ord.L -0.14756    0.10703  -1.379    0.168
toneBis.ord.Q -0.12943    0.13070  -0.990    0.322
toneBis.ord.C  0.21326    0.15164   1.406    0.160

Approximate significance of smooth terms:
                                 edf Ref.df      F  p-value    
s(measurement.no)              5.834  6.242 31.952  < 2e-16 ***
s(measurement.no):toneBis.ord1 3.602  4.282  2.215  0.06010 .  
s(measurement.no):toneBis.ord2 3.974  4.973  3.730  0.00268 ** 
s(measurement.no):toneBis.ord3 3.817  4.801  5.479 9.16e-05 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.644   Deviance explained = 67.8%
fREML = 7619.6  Scale est. = 0.2496    n = 12023

gam.check(gamm.model1e)

Method: fREML   Optimizer: perf chol
$grad
 [1] -2.606804e-13  1.605382e-13  1.389999e-13 -1.179057e-13 -3.517187e-13 -1.443290e-15 -8.105354e-05 -6.614710e-05 -8.881784e-16
[10] -3.952394e-14 -1.332268e-15 -4.180206e-05  1.776357e-14 -7.460699e-14  1.278977e-13 -5.773160e-15  3.819167e-14  6.217249e-14
[19]  1.003642e-13 -6.572520e-14  2.060574e-13  1.509903e-13  5.506706e-14  9.769963e-15 -2.398082e-14  1.332268e-14 -5.679150e-05
[28] -3.996803e-15 -2.415845e-13 -1.421085e-14  7.105427e-15  8.526513e-14 -6.394885e-14 -2.486900e-13 -1.000444e-11

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   2.263366e+00  2.957118e-02  1.182119e-01  9.913573e-03  8.148782e-02 -1.276238e-04  1.972664e-06  4.731360e-08  7.932566e-03
   2.957118e-02  7.556011e-01 -1.898898e-02  1.722273e-02  2.271054e-02 -1.042046e-05  3.166252e-05  7.383003e-09 -3.436494e-03
   1.182119e-01 -1.898898e-02  9.700982e-01 -1.434248e-02  4.049020e-02  3.774792e-04 -1.477241e-06  2.300870e-08  4.167172e-02
   9.913573e-03  1.722273e-02 -1.434248e-02  1.436278e+00 -6.220648e-03  7.418559e-05  3.627433e-07 -4.748906e-08 -1.839662e-04
   8.148782e-02  2.271054e-02  4.049020e-02 -6.220648e-03  1.844286e+01  1.261228e-02  3.470008e-06  3.777600e-06 -1.619484e-02
  -1.276238e-04 -1.042046e-05  3.774792e-04  7.418559e-05  1.261228e-02  7.502886e-02  1.730977e-08 -5.203129e-06  2.064908e-03
   1.972664e-06  3.166252e-05 -1.477241e-06  3.627433e-07  3.470008e-06  1.730977e-08  8.109396e-05 -7.505232e-11 -8.925853e-07
   4.731360e-08  7.383003e-09  2.300870e-08 -4.748906e-08  3.777600e-06 -5.203129e-06 -7.505232e-11  6.631013e-05  2.074066e-07
   7.932566e-03 -3.436494e-03  4.167172e-02 -1.839662e-04 -1.619484e-02  2.064908e-03 -8.925853e-07  2.074066e-07  4.761843e-01
  -3.208634e-05  1.544205e-06 -1.768669e-04  2.129535e-04 -9.384107e-04 -8.768483e-03  5.597922e-08 -7.209766e-10  2.331933e-02
  -8.919282e-05  2.529139e-04 -1.337485e-04  3.020934e-03 -5.203518e-02  3.208857e-04  1.692696e-07  3.174341e-08  5.633977e-03
   1.546043e-09 -4.371969e-12 -1.519128e-09 -1.163573e-08 -6.480568e-08 -2.003226e-07  5.747714e-12  1.251242e-10 -4.236830e-09
   2.212247e-02  1.437503e-02  2.249300e-02  9.336702e-03  1.246187e+00  6.126772e-04 -1.356552e-06  2.941997e-07  4.478265e-02
  -1.302330e-04 -7.000967e-05  8.955518e-04  4.041613e-04  3.621857e-02  1.462235e-01 -1.849386e-07 -1.948815e-05  6.544929e-03
   7.519140e-03  1.351893e-01 -7.072659e-03  4.244809e-03 -4.222554e-02 -2.643247e-04  3.240483e-04  2.273652e-06  6.789523e-03
   7.797280e-05  1.234992e-05  5.856251e-06 -7.068765e-05  4.079268e-03 -6.515185e-03  2.787486e-07  2.143352e-04  2.392762e-04
   3.984341e-03 -1.633067e-03  2.185301e-02  2.331700e-03 -2.147481e-02  1.007257e-03 -8.017777e-08  9.375705e-08  2.170079e-01
  -1.374653e-04  3.222476e-05  1.579074e-04  9.679396e-04  1.091004e-02  1.330298e-02 -5.852608e-09 -9.865854e-06  6.795938e-03
  -5.848222e-03 -1.153444e-04  5.310353e-03 -3.857425e-02 -1.139522e-01  3.260730e-04  6.400358e-07  1.527373e-07  4.859213e-03
  -5.413710e-05 -4.715854e-06  2.349160e-04 -2.362591e-03  1.068723e-03 -1.504986e-03  9.660242e-08  8.997901e-06  5.616847e-04
   2.409590e-02 -4.177535e-02 -3.492145e-02  1.887236e-02  2.087479e+00  1.993046e-03 -5.463874e-07  7.097418e-07 -7.890415e-03
  -7.494510e-04 -6.317616e-05  1.018112e-03  8.357567e-04  3.389941e-02  1.994890e-01  2.612850e-07 -2.172784e-05  2.787263e-03
   8.169953e-03  9.687412e-02 -3.136806e-03  2.178719e-04 -1.183260e-02  2.769589e-04  1.878768e-04  3.401809e-06  2.631574e-03
   7.144633e-05 -9.823935e-05 -9.255932e-06 -1.075036e-05  5.219847e-03 -6.619231e-03  2.165689e-07  3.633024e-04  5.253870e-04
   1.006995e-04  4.634144e-04  2.530239e-04  6.782945e-04 -3.924636e-02  4.410851e-04  2.115793e-07  1.837322e-08  1.071768e-01
   5.949262e-05  8.428063e-07 -1.664917e-05 -6.792660e-05  7.375301e-04  9.490443e-04 -1.936110e-08  4.156107e-06  7.534744e-04
   4.754224e-08  6.684307e-09  5.747741e-08  5.081872e-07 -3.838214e-06  2.281734e-08 -2.759643e-11 -1.107925e-12  2.643940e-07
  -2.932836e-05  6.988533e-06 -1.963514e-05  5.319037e-04 -2.692417e-04 -2.239010e-03 -4.808137e-08 -2.471294e-06  2.727444e-04
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -3.208634e-05 -8.919282e-05  1.546043e-09  2.212247e-02 -1.302330e-04  7.519140e-03  7.797280e-05  3.984341e-03 -1.374653e-04
   1.544205e-06  2.529139e-04 -4.371969e-12  1.437503e-02 -7.000967e-05  1.351893e-01  1.234992e-05 -1.633067e-03  3.222476e-05
  -1.768669e-04 -1.337485e-04 -1.519128e-09  2.249300e-02  8.955518e-04 -7.072659e-03  5.856251e-06  2.185301e-02  1.579074e-04
   2.129535e-04  3.020934e-03 -1.163573e-08  9.336702e-03  4.041613e-04  4.244809e-03 -7.068765e-05  2.331700e-03  9.679396e-04
  -9.384107e-04 -5.203518e-02 -6.480568e-08  1.246187e+00  3.621857e-02 -4.222554e-02  4.079268e-03 -2.147481e-02  1.091004e-02
  -8.768483e-03  3.208857e-04 -2.003226e-07  6.126772e-04  1.462235e-01 -2.643247e-04 -6.515185e-03  1.007257e-03  1.330298e-02
   5.597922e-08  1.692696e-07  5.747714e-12 -1.356552e-06 -1.849386e-07  3.240483e-04  2.787486e-07 -8.017777e-08 -5.852608e-09
  -7.209766e-10  3.174341e-08  1.251242e-10  2.941997e-07 -1.948815e-05  2.273652e-06  2.143352e-04  9.375705e-08 -9.865854e-06
   2.331933e-02  5.633977e-03 -4.236830e-09  4.478265e-02  6.544929e-03  6.789523e-03  2.392762e-04  2.170079e-01  6.795938e-03
   1.844055e+00 -6.659143e-05 -1.033513e-06  4.462054e-04  2.475496e-02  5.250636e-04  2.043876e-03  8.314700e-03  4.001232e-01
  -6.659143e-05  7.378615e-02  3.971181e-07  4.726695e-03  1.061047e-03 -1.981501e-03  6.919535e-05  1.547852e-03  1.075918e-04
  -1.033513e-06  3.971181e-07  4.180454e-05 -6.001818e-08  1.439146e-06  1.822149e-08  1.548986e-07  1.203567e-10 -5.729976e-07
   4.462054e-04  4.726695e-03 -6.001818e-08  6.078054e+00  2.240296e-02 -2.403236e-01  5.130353e-04 -3.170258e-02 -5.296697e-03
   2.475496e-02  1.061047e-03  1.439146e-06  2.240296e-02  9.390697e+00 -3.608698e-03  1.510097e-02  2.752722e-03  7.663765e-01
   5.250636e-04 -1.981501e-03  1.822149e-08 -2.403236e-01 -3.608698e-03  1.170472e+01  7.623010e-02  1.003060e-02  6.521751e-03
   2.043876e-03  6.919535e-05  1.548986e-07  5.130353e-04  1.510097e-02  7.623010e-02  1.022238e+00  1.008682e-04 -1.680528e-02
   8.314700e-03  1.547852e-03  1.203567e-10 -3.170258e-02  2.752722e-03  1.003060e-02  1.008682e-04  5.362080e-01  9.192892e-03
   4.001232e-01  1.075918e-04 -5.729976e-07 -5.296697e-03  7.663765e-01  6.521751e-03 -1.680528e-02  9.192892e-03  9.239379e+00
   2.025290e-05  9.804807e-02  5.296315e-07  9.742284e-02 -7.699768e-04  7.732701e-03  2.635856e-04  8.148788e-03  1.228302e-03
  -4.604084e-03  7.676257e-03  6.738220e-05  1.441764e-04 -1.296469e-01  4.027768e-03  3.607445e-02  8.054612e-04 -4.221971e-03
  -4.194604e-04 -2.364254e-02 -5.145997e-08  1.285753e-01  7.724224e-03 -8.421751e-02  8.143996e-04 -6.345837e-02 -6.725808e-04
  -1.603306e-02  8.390177e-04 -1.504720e-06 -7.962855e-04  5.432809e-01  7.068289e-05 -6.072112e-03  1.245527e-03 -3.151877e-02
   5.955693e-04 -5.548982e-04  2.977991e-08 -6.762885e-02  5.447921e-04  2.924078e+00  1.001372e-02  3.458527e-03  7.607383e-04
   6.611094e-03  7.766502e-05  1.309469e-07 -2.651135e-05  5.626045e-03  3.856762e-02  5.778893e-01  2.278884e-04 -4.198671e-03
   4.588354e-03  1.560705e-03 -7.776654e-09  1.547766e-02  1.676971e-03  8.943004e-03  3.520276e-05  4.236934e-02  1.065990e-03
   8.731027e-02  1.654048e-04 -1.159028e-06  1.123214e-03 -5.138385e-03  2.837903e-04  1.025614e-02 -8.378921e-04 -2.693567e-01
  -1.838551e-08  4.720439e-06  9.410561e-12  7.046007e-08  1.581500e-07 -5.490081e-07 -4.864264e-10  1.434818e-07  2.678472e-09
  -1.034248e-03  2.346838e-03  1.229876e-05 -1.080354e-03 -1.755329e-02 -7.496239e-04 -5.288512e-03  1.223291e-04  4.510787e-03
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
  -5.848222e-03 -5.413710e-05  2.409590e-02 -7.494510e-04  8.169953e-03  7.144633e-05  1.006995e-04  5.949262e-05  4.754224e-08
  -1.153444e-04 -4.715854e-06 -4.177535e-02 -6.317616e-05  9.687412e-02 -9.823935e-05  4.634144e-04  8.428063e-07  6.684307e-09
   5.310353e-03  2.349160e-04 -3.492145e-02  1.018112e-03 -3.136806e-03 -9.255932e-06  2.530239e-04 -1.664917e-05  5.747741e-08
  -3.857425e-02 -2.362591e-03  1.887236e-02  8.357567e-04  2.178719e-04 -1.075036e-05  6.782945e-04 -6.792660e-05  5.081872e-07
  -1.139522e-01  1.068723e-03  2.087479e+00  3.389941e-02 -1.183260e-02  5.219847e-03 -3.924636e-02  7.375301e-04 -3.838214e-06
   3.260730e-04 -1.504986e-03  1.993046e-03  1.994890e-01  2.769589e-04 -6.619231e-03  4.410851e-04  9.490443e-04  2.281734e-08
   6.400358e-07  9.660242e-08 -5.463874e-07  2.612850e-07  1.878768e-04  2.165689e-07  2.115793e-07 -1.936110e-08 -2.759643e-11
   1.527373e-07  8.997901e-06  7.097418e-07 -2.172784e-05  3.401809e-06  3.633024e-04  1.837322e-08  4.156107e-06 -1.107925e-12
   4.859213e-03  5.616847e-04 -7.890415e-03  2.787263e-03  2.631574e-03  5.253870e-04  1.071768e-01  7.534744e-04  2.643940e-07
   2.025290e-05 -4.604084e-03 -4.194604e-04 -1.603306e-02  5.955693e-04  6.611094e-03  4.588354e-03  8.731027e-02 -1.838551e-08
   9.804807e-02  7.676257e-03 -2.364254e-02  8.390177e-04 -5.548982e-04  7.766502e-05  1.560705e-03  1.654048e-04  4.720439e-06
   5.296315e-07  6.738220e-05 -5.145997e-08 -1.504720e-06  2.977991e-08  1.309469e-07 -7.776654e-09 -1.159028e-06  9.410561e-12
   9.742284e-02  1.441764e-04  1.285753e-01 -7.962855e-04 -6.762885e-02 -2.651135e-05  1.547766e-02  1.123214e-03  7.046007e-08
  -7.699768e-04 -1.296469e-01  7.724224e-03  5.432809e-01  5.447921e-04  5.626045e-03  1.676971e-03 -5.138385e-03  1.581500e-07
   7.732701e-03  4.027768e-03 -8.421751e-02  7.068289e-05  2.924078e+00  3.856762e-02  8.943004e-03  2.837903e-04 -5.490081e-07
   2.635856e-04  3.607445e-02  8.143996e-04 -6.072112e-03  1.001372e-02  5.778893e-01  3.520276e-05  1.025614e-02 -4.864264e-10
   8.148788e-03  8.054612e-04 -6.345837e-02  1.245527e-03  3.458527e-03  2.278884e-04  4.236934e-02 -8.378921e-04  1.434818e-07
   1.228302e-03 -4.221971e-03 -6.725808e-04 -3.151877e-02  7.607383e-04 -4.198671e-03  1.065990e-03 -2.693567e-01  2.678472e-09
   4.974034e-01  3.558081e-02 -7.606287e-02 -3.933948e-04  1.760611e-03  3.988876e-04  5.417424e-03  2.710934e-04  8.884453e-06
   3.558081e-02  7.498610e+00 -4.850523e-03 -8.377111e-02  1.383524e-03  4.102462e-02  7.405094e-05 -5.196110e-02  3.733020e-07
  -7.606287e-02 -4.850523e-03  8.671335e+00  1.539550e-01  5.091225e-02  1.139065e-03  1.006212e-01  7.291063e-05 -8.280757e-06
  -3.933948e-04 -8.377111e-02  1.539550e-01  6.386262e+00  1.030344e-03 -5.966447e-03  8.445894e-04  2.009516e-01  1.081214e-07
   1.760611e-03  1.383524e-03  5.091225e-02  1.030344e-03  3.105909e+00  2.404869e-02  4.697770e-03  1.310977e-03  1.955576e-07
   3.988876e-04  4.102462e-02  1.139065e-03 -5.966447e-03  2.404869e-02  2.437535e+00  2.347651e-06 -1.377986e-02  2.986450e-09
   5.417424e-03  7.405094e-05  1.006212e-01  8.445894e-04  4.697770e-03  2.347651e-06  7.824234e-02 -2.463124e-03  2.532264e-09
   2.710934e-04 -5.196110e-02  7.291063e-05  2.009516e-01  1.310977e-03 -1.377986e-02 -2.463124e-03  9.800043e-01  9.447840e-08
   8.884453e-06  3.733020e-07 -8.280757e-06  1.081214e-07  1.955576e-07  2.986450e-09  2.532264e-09  9.447840e-08  5.679233e-05
   3.630302e-03  3.406304e-01 -5.064643e-04 -1.443359e-01 -6.578235e-04 -1.406511e-02  1.661884e-04  6.426981e-03 -7.383096e-07
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -2.932836e-05 -1.728620e-02  4.915948e-04  1.370490e-03  1.316025e-04 -5.051627e-03  8.305357e-04 -2.416918e+00
   6.988533e-06  8.168352e-02  1.910039e-04 -6.318119e-03  9.413799e-05  3.006954e-02  1.171627e-04 -1.300793e+00
  -1.963514e-05 -8.799556e-02 -2.550957e-04  3.200506e-03  2.191247e-04  1.599280e-02 -3.295337e-04 -1.487161e+00
   5.319037e-04 -1.277637e-02 -1.657296e-03  2.317459e-03  5.899133e-04  5.797676e-02 -3.327210e-03 -1.408451e+00
  -2.692417e-04 -1.001046e-01  3.303645e-02 -1.043963e-01  1.422508e-02  1.954781e-03  3.931593e-02 -2.675274e+01
  -2.239010e-03  2.146999e-03 -2.235130e-02  8.182145e-04  1.216603e-03  6.435757e-04  1.673685e-02 -5.789188e-01
  -4.808137e-08 -1.087285e-05 -5.361425e-08 -8.365357e-06  6.402595e-07  8.433642e-06 -3.232011e-07 -1.191399e-03
  -2.471294e-06 -6.434546e-07  3.371109e-05 -1.912569e-07 -1.183281e-04 -5.110674e-07  1.567518e-05 -9.321882e-04
   2.727444e-04  9.922870e-02  3.815046e-03  1.851736e-02  2.644450e-03  9.815461e-02  2.670743e-03 -3.302436e+00
  -1.034248e-03 -3.393346e-03  1.217942e-02 -1.232550e-03  1.666439e-02 -4.616135e-03  9.570525e-03 -2.842589e+00
   2.346838e-03 -4.421501e-03  5.259171e-04  1.972192e-02 -4.236840e-04  1.035805e-01  2.790196e-05 -1.666341e+00
   1.229876e-05  8.627996e-07  3.494815e-06 -4.955435e-08 -6.269002e-06  1.162362e-07 -1.704904e-05 -1.400779e-04
  -1.080354e-03  8.523022e-01 -5.418694e-03  1.200342e-01 -6.531862e-05  3.210475e-01 -1.331550e-02 -2.401139e+01
  -1.755329e-02 -1.614168e-02  5.749038e-02 -2.735649e-03  9.304945e-02 -3.542529e-02  8.668604e-01 -1.717365e+01
  -7.496239e-04  5.871420e-03 -1.236314e-03  1.476181e-01  4.579091e-02  2.388808e-01 -2.330675e-02 -3.321135e+01
  -5.288512e-03 -5.736584e-03  6.308972e-02  6.577524e-04  1.576013e-02 -2.292191e-04 -6.249886e-03 -3.112460e+00
   1.223291e-04  9.500309e-02  1.633004e-03  6.021000e-02 -8.854823e-04  9.916253e-02 -2.375200e-03 -5.802650e+00
   4.510787e-03 -4.914047e-03 -2.821604e-01 -9.518177e-03 -8.187934e-01 -1.856565e-03 -1.124877e+00 -1.359522e+01
   3.630302e-03  1.137339e-01  9.762798e-04  1.360212e-02 -2.260512e-03  3.031421e-01 -1.426618e-03 -6.022106e+00
   3.406304e-01  6.118956e-02 -3.366006e-01  1.060161e-02 -4.538047e-01  2.528679e-02 -6.929985e-01 -1.018046e+01
  -5.064643e-04  1.214331e-01  4.790491e-03 -2.586071e-01 -1.231978e-03 -7.157577e-01  8.633410e-03 -2.436597e+01
  -1.443359e-01 -4.120117e-03 -7.048702e-02 -6.091438e-04 -3.171609e-01 -1.716519e-03 -1.016692e-01 -9.056666e+00
  -6.578235e-04 -2.717570e-01  1.525364e-03  4.896323e-02  1.010079e-02 -1.173263e-01  3.172594e-03 -1.386230e+01
  -1.406511e-02 -6.206100e-03 -1.257657e-01  7.147744e-05 -6.704679e-02 -1.175774e-03 -1.037371e-01 -4.799202e+00
   1.661884e-04  2.759373e-02  1.207553e-03  2.756418e-02 -1.943155e-03  1.819321e-02  1.061517e-03 -2.147538e+00
   6.426981e-03 -6.183219e-03 -1.474789e-01  2.449890e-04 -1.942918e-01  4.109374e-03 -6.494298e-02 -3.323144e+00
  -7.383096e-07  5.122037e-06 -3.211159e-08  1.329951e-06  2.572285e-07  1.021388e-05 -8.907799e-08 -2.631901e-04
   6.310079e-01 -6.257743e-03  3.958461e-02 -5.718670e-03  2.328244e-01  2.149611e-03 -2.959794e-02 -2.485607e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  6565 / 6565 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 5.83e+00    0.99    0.19
s(measurement.no):toneBis.ord1                      9.00e+00 3.60e+00    0.99    0.23
s(measurement.no):toneBis.ord2                      9.00e+00 3.97e+00    0.99    0.17
s(measurement.no):toneBis.ord3                      9.00e+00 3.82e+00    0.99    0.21
s(measurement.no,speaker)                           1.00e+02 5.47e+01    0.99    0.24
s(measurement.no,speaker):toneBis.ord1              1.00e+02 4.54e-03    0.99    0.21
s(measurement.no,speaker):toneBis.ord2              1.00e+02 1.23e+01    0.99    0.20
s(measurement.no,speaker):toneBis.ord3              1.00e+02 3.33e+00    0.99    0.23
s(measurement.no,speakerLeftRightTone)              5.90e+02 8.24e+01    0.99    0.18
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.90e+02 7.26e+01    0.99    0.24
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.90e+02 3.88e+01    0.99    0.17
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.90e+02 3.24e+01    0.99    0.23
s(measurement.no,speakerPos)                        3.00e+02 6.68e+01    0.99    0.17
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 3.73e+01    0.99    0.20
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 1.09e+01    0.99    0.20
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 4.97e+00    0.99    0.21
s(measurement.no,word)                              6.30e+02 3.10e+02    0.99    0.25
s(measurement.no,wordPos)                           9.99e+02 1.70e+02    0.99    0.23
s(measurement.no,wordLeftRightTone)                 9.36e+02 2.27e+02    0.99    0.18

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1e, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F2 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1e, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1e, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1e, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1e, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(1400, 1800), xlab = "Time (ms)", ylab = "F2 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1 * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1e, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2 * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1e, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3 * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1e, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4 * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("bottomright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1e, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 2.020202

plot_diff(gamm.model1e, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 1.717172

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1e, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1e, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1e, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 1.919192

plot_diff(gamm.model1e, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 1.414141

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F2 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1e, pred = list(toneBis.ord=c("1", "2", "3", "4")), xlab = "F2 (Z)", main = "Tone")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.7 Model 1F: /ai/ female, F2 as output

gamm.model1f.noAR <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                        
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1f.noAR, paste("Gamm_model1f_noAR.rds"))

gamm.model1f.noAR <- 
  readRDS("Gamm_model1f_noAR.rds")

r.gamm.model1f <- start_value_rho(gamm.model1f.noAR)

# Auto-regressive model

gamm.model1f <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                          
                        
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                    
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                    
                      data=data.ai.fem, method="fREML", rho = r.gamm.model1f, AR.start = data.ai.fem$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1f, paste("Gamm_model1f.rds"))

gamm.model1f <- 
  readRDS("Gamm_model1f.rds")

summary(gamm.model1f, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept)   0.01759    0.06436   0.273    0.785
toneBis.ord1 -0.10687    0.24338  -0.439    0.661
toneBis.ord2  0.07739    0.12629   0.613    0.540
toneBis.ord3 -0.02991    0.16311  -0.183    0.854

Approximate significance of smooth terms:
                                 edf Ref.df     F  p-value    
s(measurement.no)              2.993  3.301 8.992 3.41e-06 ***
s(measurement.no):toneBis.ord1 1.698  1.920 1.032   0.2849    
s(measurement.no):toneBis.ord2 1.000  1.001 3.952   0.0468 *  
s(measurement.no):toneBis.ord3 1.262  1.360 1.279   0.2068    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.469   Deviance explained = 51.9%
fREML =  10968  Scale est. = 0.45463   n = 11198

gam.check(gamm.model1f)

Method: fREML   Optimizer: perf chol
$grad
 [1] -8.680216e-09 -3.818379e-08 -3.832142e-05 -2.706223e-07  2.075530e-06 -4.120451e-09 -5.147328e-05 -4.062929e-05 -4.213564e-05
[10] -1.307415e-08 -3.909493e-05  1.569795e-07 -4.827928e-06  1.250390e-07 -2.757438e-08 -2.751360e-08 -1.294878e-08  5.329621e-08
[19] -4.373979e-05 -3.576229e-07 -3.297578e-07  5.353744e-10 -7.308002e-08 -1.399523e-08 -2.017717e-08 -3.045525e-09 -8.080303e-07
[28]  1.043800e-07 -2.594111e-06 -1.904703e-07 -1.472385e-05  5.402271e-07 -2.314541e-06 -2.590902e-07  2.108991e-04

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   5.217266e-01 -1.930705e-02 -3.226213e-05  8.397579e-03  3.502305e-01 -2.819300e-04 -3.039013e-07  7.613186e-10 -1.109815e-06
  -1.930705e-02  1.520758e-01 -6.195473e-06  3.091282e-03  7.003164e-04 -4.871926e-04  3.531726e-07  4.822612e-09 -1.757666e-07
  -3.226213e-05 -6.195473e-06  3.833632e-05  4.997395e-06  5.900655e-06 -7.992972e-08 -6.589872e-11  2.940762e-13  6.031342e-10
   8.397579e-03  3.091282e-03  4.997395e-06  1.477972e-02 -1.743697e-02  1.406031e-04  2.625273e-08  1.728462e-09  1.631168e-07
   3.502305e-01  7.003164e-04  5.900655e-06 -1.743697e-02  1.860259e+01  3.349125e-02  2.994159e-06  2.667247e-07  2.023175e-05
  -2.819300e-04 -4.871926e-04 -7.992972e-08  1.406031e-04  3.349125e-02  2.801516e-01 -2.402172e-08 -6.602727e-07 -8.595425e-08
  -3.039013e-07  3.531726e-07 -6.589872e-11  2.625273e-08  2.994159e-06 -2.402172e-08  5.147416e-05  5.204887e-13 -9.187162e-12
   7.613186e-10  4.822612e-09  2.940762e-13  1.728462e-09  2.667247e-07 -6.602727e-07  5.204887e-13  4.062969e-05  6.404318e-13
  -1.109815e-06 -1.757666e-07  6.031342e-10  1.631168e-07  2.023175e-05 -8.595425e-08 -9.187162e-12  6.404318e-13  4.213742e-05
   2.165835e-05  3.484902e-05 -2.463351e-07  1.945314e-04  2.555953e-03 -4.144252e-02  2.336192e-09 -1.868238e-08 -7.766648e-07
   2.321810e-02  9.882114e-03  6.855199e-06  4.347112e-02 -5.462106e-01 -2.605900e-05  3.524206e-07 -1.833161e-08  1.799308e-06
  -1.617145e-04 -1.543894e-04 -3.595463e-08 -4.685308e-04  9.308661e-07 -3.650114e-02 -5.786495e-10 -4.171554e-07  3.279861e-08
   1.646267e-02  3.734402e-03  4.332693e-06 -4.723268e-04  7.219024e-01  9.611287e-04  1.184157e-06  1.596543e-08 -9.358312e-08
   3.393494e-04 -7.691454e-04  6.592938e-08  2.845870e-04  4.839794e-02  3.149861e-01 -3.388703e-08  1.019625e-06 -4.013356e-08
  -2.752228e-03  7.911465e-03 -6.700675e-07  2.698351e-04  2.883183e-02 -1.590480e-04  6.650317e-06 -1.218173e-07 -1.422349e-07
   2.302906e-05 -1.069552e-04 -1.736894e-08  4.606923e-05  1.415211e-03 -1.258425e-02  9.769980e-08 -2.199708e-06  7.523372e-09
  -5.445899e-03 -1.462877e-03  5.236577e-06 -7.837374e-05  9.707050e-02  3.413423e-04 -1.584488e-07  3.951092e-09  6.409136e-06
   4.630002e-04  2.028613e-04  1.057648e-07 -2.348404e-04 -1.082255e-02 -2.751171e-02  3.961053e-09 -8.798374e-08 -2.259283e-07
   8.681042e-09  3.818900e-08  8.116081e-11  2.706613e-07 -2.075839e-06  4.120654e-09  3.189484e-12  1.324348e-13 -1.499435e-11
  -4.795063e-05 -1.191310e-05  8.050279e-08 -7.326568e-05 -2.711915e-03 -5.484909e-03  4.460416e-09  2.861714e-09  3.504736e-09
   1.607131e-02 -1.545276e-03  7.384527e-07 -2.901245e-03  1.083799e+00  2.306361e-03 -1.522510e-07  2.382708e-08  3.401256e-06
  -4.780234e-05 -7.913123e-04 -1.812025e-07  1.439888e-04  2.615313e-02  2.091210e-01 -2.684044e-08 -7.233151e-07 -7.321848e-09
  -7.070520e-03 -1.624012e-02 -2.660129e-06  1.206649e-03  2.547314e-01 -1.610947e-03  5.395028e-05  3.727995e-07 -6.899120e-07
  -2.325632e-04  1.614707e-03  2.176382e-08  9.042713e-05  4.385212e-03 -1.488950e-02  1.281461e-07 -1.535037e-06  3.014524e-08
  -1.359850e-02 -3.804345e-03  1.142487e-05  4.752608e-03  3.439529e-01 -8.074640e-04 -2.795746e-07  1.736802e-08  2.952453e-05
  -5.751961e-05  1.297463e-05 -4.042180e-07  2.297452e-04 -1.115517e-03 -2.266220e-02 -1.965011e-09 -2.391081e-07 -2.670192e-07
   4.605224e-04  1.321624e-04  4.877003e-08  5.598466e-04 -2.588102e-03  1.596977e-05  4.000689e-09  6.003146e-10  8.981157e-08
  -9.803994e-05 -3.677274e-05 -5.865504e-08 -1.081200e-04 -1.143095e-03 -1.700310e-02 -1.658247e-09 -1.941430e-07 -1.955776e-08
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   2.165835e-05  2.321810e-02 -1.617145e-04  1.646267e-02  3.393494e-04 -2.752228e-03  2.302906e-05 -5.445899e-03  4.630002e-04
   3.484902e-05  9.882114e-03 -1.543894e-04  3.734402e-03 -7.691454e-04  7.911465e-03 -1.069552e-04 -1.462877e-03  2.028613e-04
  -2.463351e-07  6.855199e-06 -3.595463e-08  4.332693e-06  6.592938e-08 -6.700675e-07 -1.736894e-08  5.236577e-06  1.057648e-07
   1.945314e-04  4.347112e-02 -4.685308e-04 -4.723268e-04  2.845870e-04  2.698351e-04  4.606923e-05 -7.837374e-05 -2.348404e-04
   2.555953e-03 -5.462106e-01  9.308661e-07  7.219024e-01  4.839794e-02  2.883183e-02  1.415211e-03  9.707050e-02 -1.082255e-02
  -4.144252e-02 -2.605900e-05 -3.650114e-02  9.611287e-04  3.149861e-01 -1.590480e-04 -1.258425e-02  3.413423e-04 -2.751171e-02
   2.336192e-09  3.524206e-07 -5.786495e-10  1.184157e-06 -3.388703e-08  6.650317e-06  9.769980e-08 -1.584488e-07  3.961053e-09
  -1.868238e-08 -1.833161e-08 -4.171554e-07  1.596543e-08  1.019625e-06 -1.218173e-07 -2.199708e-06  3.951092e-09 -8.798374e-08
  -7.766648e-07  1.799308e-06  3.279861e-08 -9.358312e-08 -4.013356e-08 -1.422349e-07  7.523372e-09  6.409136e-06 -2.259283e-07
   9.569397e-01  3.284336e-04 -1.491467e-02  7.817115e-04 -5.723585e-03  4.667344e-05 -2.432069e-03 -3.297880e-03  3.681986e-01
   3.284336e-04  1.007836e+01 -2.876165e-02  1.053684e-01 -8.568196e-03  3.971283e-03  4.233643e-04  3.836047e-02  9.079782e-04
  -1.491467e-02 -2.876165e-02  1.482623e+00 -8.420342e-04  4.928133e-02  1.093213e-04  6.794024e-04  3.634853e-04 -1.015902e-02
   7.817115e-04  1.053684e-01 -8.420342e-04  2.156814e+00 -3.114549e-02  1.160420e-02 -1.019431e-03 -5.563219e-02  8.939109e-04
  -5.723585e-03 -8.568196e-03  4.928133e-02 -3.114549e-02  6.436357e+00 -4.796883e-04 -2.836515e-02  1.260381e-03 -1.086346e-01
   4.667344e-05  3.971283e-03  1.093213e-04  1.160420e-02 -4.796883e-04  4.728008e-01 -4.813332e-02  6.629405e-04 -4.612784e-04
  -2.432069e-03  4.233643e-04  6.794024e-04 -1.019431e-03 -2.836515e-02 -4.813332e-02  3.055625e+00 -4.440650e-04  2.514453e-02
  -3.297880e-03  3.836047e-02  3.634853e-04 -5.563219e-02  1.260381e-03  6.629405e-04 -4.440650e-04  5.873026e-01 -3.240744e-02
   3.681986e-01  9.079782e-04 -1.015902e-02  8.939109e-04 -1.086346e-01 -4.612784e-04  2.514453e-02 -3.240744e-02  4.660306e+00
   1.307617e-08  3.910114e-05 -1.570021e-07  4.828856e-06 -1.250639e-07  2.758058e-08  2.751715e-08  1.294204e-08 -5.330293e-08
   2.459841e-03  2.825781e-03  3.063558e-01  2.109926e-03 -1.510518e-02 -7.913843e-05  2.632136e-02  5.192854e-04 -3.424336e-02
  -3.052525e-06  7.121528e-02  5.511540e-04 -5.782026e-02 -1.573032e-03 -7.583509e-03  3.624428e-04  3.272084e-02 -9.663715e-04
  -1.490678e-03  1.471578e-03 -3.916288e-03  2.558795e-03  3.641941e-01  3.436058e-04 -6.882789e-02  4.517116e-04  2.558457e-02
   2.748291e-04  5.797480e-03  1.379381e-04  9.859597e-02  2.057709e-03  7.779636e-01  1.846493e-03 -9.492715e-03  1.023744e-03
   5.747295e-05 -6.968471e-04 -1.005144e-02  1.908172e-03  4.606862e-02 -4.819446e-03  1.629859e-01  2.717590e-04 -1.016687e-04
  -1.182489e-02  1.437834e-02  2.174949e-04  1.072130e-01  3.257204e-03 -2.209965e-03  2.372833e-04  1.779642e-01 -9.030754e-03
   3.058719e-01  1.314776e-03 -1.231566e-02 -1.377482e-03 -1.364549e-01 -6.217936e-06 -3.439337e-03 -2.515359e-03  4.955732e-02
  -7.205593e-06  9.115598e-02  1.198224e-05  5.523629e-03 -1.980450e-04  4.086530e-06  8.968962e-06  1.173718e-03 -1.707707e-04
   1.631130e-03 -1.706391e-02  5.107935e-01 -1.918595e-03  2.283547e-03  9.244644e-05  2.655193e-03 -2.547562e-04  2.955204e-02
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   8.681042e-09 -4.795063e-05  1.607131e-02 -4.780234e-05 -7.070520e-03 -2.325632e-04 -1.359850e-02 -5.751961e-05  4.605224e-04
   3.818900e-08 -1.191310e-05 -1.545276e-03 -7.913123e-04 -1.624012e-02  1.614707e-03 -3.804345e-03  1.297463e-05  1.321624e-04
   8.116081e-11  8.050279e-08  7.384527e-07 -1.812025e-07 -2.660129e-06  2.176382e-08  1.142487e-05 -4.042180e-07  4.877003e-08
   2.706613e-07 -7.326568e-05 -2.901245e-03  1.439888e-04  1.206649e-03  9.042713e-05  4.752608e-03  2.297452e-04  5.598466e-04
  -2.075839e-06 -2.711915e-03  1.083799e+00  2.615313e-02  2.547314e-01  4.385212e-03  3.439529e-01 -1.115517e-03 -2.588102e-03
   4.120654e-09 -5.484909e-03  2.306361e-03  2.091210e-01 -1.610947e-03 -1.488950e-02 -8.074640e-04 -2.266220e-02  1.596977e-05
   3.189484e-12  4.460416e-09 -1.522510e-07 -2.684044e-08  5.395028e-05  1.281461e-07 -2.795746e-07 -1.965011e-09  4.000689e-09
   1.324348e-13  2.861714e-09  2.382708e-08 -7.233151e-07  3.727995e-07 -1.535037e-06  1.736802e-08 -2.391081e-07  6.003146e-10
  -1.499435e-11  3.504736e-09  3.401256e-06 -7.321848e-09 -6.899120e-07  3.014524e-08  2.952453e-05 -2.670192e-07  8.981157e-08
   1.307617e-08  2.459841e-03 -3.052525e-06 -1.490678e-03  2.748291e-04  5.747295e-05 -1.182489e-02  3.058719e-01 -7.205593e-06
   3.910114e-05  2.825781e-03  7.121528e-02  1.471578e-03  5.797480e-03 -6.968471e-04  1.437834e-02  1.314776e-03  9.115598e-02
  -1.570021e-07  3.063558e-01  5.511540e-04 -3.916288e-03  1.379381e-04 -1.005144e-02  2.174949e-04 -1.231566e-02  1.198224e-05
   4.828856e-06  2.109926e-03 -5.782026e-02  2.558795e-03  9.859597e-02  1.908172e-03  1.072130e-01 -1.377482e-03  5.523629e-03
  -1.250639e-07 -1.510518e-02 -1.573032e-03  3.641941e-01  2.057709e-03  4.606862e-02  3.257204e-03 -1.364549e-01 -1.980450e-04
   2.758058e-08 -7.913843e-05 -7.583509e-03  3.436058e-04  7.779636e-01 -4.819446e-03 -2.209965e-03 -6.217936e-06  4.086530e-06
   2.751715e-08  2.632136e-02  3.624428e-04 -6.882789e-02  1.846493e-03  1.629859e-01  2.372833e-04 -3.439337e-03  8.968962e-06
   1.294204e-08  5.192854e-04  3.272084e-02  4.517116e-04 -9.492715e-03  2.717590e-04  1.779642e-01 -2.515359e-03  1.173718e-03
  -5.330293e-08 -3.424336e-02 -9.663715e-04  2.558457e-02  1.023744e-03 -1.016687e-04 -9.030754e-03  4.955732e-02 -1.707707e-04
   4.374049e-05  3.576644e-07  3.298218e-07 -5.358923e-10  7.309128e-08  1.399742e-08  2.016450e-08  3.046670e-09  8.081566e-07
   3.576644e-07  8.867639e-01 -1.785381e-03 -6.090648e-03  3.013441e-05 -2.874120e-03  4.528766e-04 -1.325350e-02 -4.657195e-04
   3.298218e-07 -1.785381e-03  2.520589e+00 -8.585768e-02 -7.007399e-03  2.981017e-05  5.028011e-01 -5.296312e-03  1.441468e-03
  -5.358923e-10 -6.090648e-03 -8.585768e-02  3.310900e+00  2.033160e-03 -8.632237e-02  1.991545e-03 -1.935478e-01 -8.062306e-05
   7.309128e-08  3.013441e-05 -7.007399e-03  2.033160e-03  1.538678e+01  7.028293e-02 -2.234707e-02  2.307943e-03 -2.184207e-05
   1.399742e-08 -2.874120e-03  2.981017e-05 -8.632237e-02  7.028293e-02  3.461121e+00  1.783457e-04 -1.001350e-02  2.182243e-05
   2.016450e-08  4.528766e-04  5.028011e-01  1.991545e-03 -2.234707e-02  1.783457e-04  3.292450e+00 -4.184621e-02  6.752231e-03
   3.046670e-09 -1.325350e-02 -5.296312e-03 -1.935478e-01  2.307943e-03 -1.001350e-02 -4.184621e-02  2.467040e+00  1.432436e-04
   8.081566e-07 -4.657195e-04  1.441468e-03 -8.062306e-05 -2.184207e-05  2.182243e-05  6.752231e-03  1.432436e-04  9.515830e-03
  -1.043987e-07  1.037198e-01 -3.727596e-03 -5.587930e-02  5.552556e-04 -8.557728e-03 -3.201470e-05  1.201616e-02  1.910125e-04
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -9.803994e-05 -4.747372e-02  3.337757e-04 -5.418671e-03  4.361552e-04 -8.781344e-03  1.355612e-04 -9.964186e-01
  -3.677274e-05  4.338458e-02 -2.286561e-03  3.225563e-02 -1.657239e-03  2.674596e-02 -9.082588e-04 -3.489283e-01
  -5.865504e-08  2.123250e-05  2.014125e-06  9.459766e-06  6.666164e-07  1.840276e-05  2.585350e-07 -1.428383e-04
  -1.081200e-04  1.233268e-02  1.473303e-04  5.694499e-03  2.511116e-04 -1.249822e-03 -1.267570e-04 -1.310067e-01
  -1.143095e-03  2.602325e-01 -3.165838e-03 -1.789058e-02 -6.548626e-03 -6.975417e-02 -1.770740e-03 -2.640258e+01
  -1.700310e-02 -3.933377e-03 -6.558814e-02 -3.409374e-03 -4.227723e-02 -1.558701e-04 -6.161006e-02 -1.083381e+00
  -1.658247e-09  1.880510e-06  1.737619e-08  3.734035e-06  8.451515e-09  1.209781e-06 -4.415712e-08 -2.045015e-04
  -1.941430e-07 -9.256016e-08  3.593163e-06  1.127717e-07  4.895491e-06 -8.009378e-08 -2.099717e-06 -4.555024e-05
  -1.955776e-08  1.617998e-06  1.266259e-07 -3.290745e-06 -1.455549e-07  6.832467e-06  2.447451e-07 -2.497555e-04
   1.631130e-03 -3.313858e-03 -2.050332e-01 -7.525615e-04 -1.239484e-01  2.250413e-03 -8.866007e-02 -2.060179e+00
  -1.706391e-02 -2.347254e-01 -1.041339e-02  5.520654e-01 -9.369219e-03  1.838998e-02 -7.607017e-03 -1.685662e+01
   5.107935e-01  1.048434e-02 -1.044820e-01  3.240730e-03 -4.986571e-02  2.328300e-03 -4.790406e-02 -2.313245e+00
  -1.918595e-03 -1.566408e-01 -6.806189e-03  4.001113e-02 -8.458014e-03  3.329030e-02 -8.507843e-03 -1.455140e+01
   2.283547e-03 -1.291251e-02 -3.642201e-02 -2.424088e-03  7.826512e-02 -1.000503e-02 -6.166251e-01 -1.481612e+01
   9.244644e-05  2.332050e-02 -2.733424e-03  3.963072e-02  1.159562e-03 -4.736459e-02  2.169779e-03 -6.155394e+00
   2.655193e-03 -2.738249e-02 -1.557734e-02 -1.377524e-02  9.622281e-02 -1.082357e-02 -1.048492e-01 -5.121165e+00
  -2.547562e-04  6.985141e-02 -8.414637e-04  1.849721e-02  1.911055e-03  1.819945e-01  1.168761e-03 -6.559820e+00
   2.955204e-02 -1.490551e-02  1.263609e-01 -1.007583e-02  1.283007e-01 -2.835415e-02 -2.762107e-02 -1.055743e+01
  -1.043987e-07  2.594697e-06  1.905009e-07  1.472650e-05 -5.403156e-07  2.315052e-06  2.591308e-07 -2.109292e-04
   1.037198e-01  9.852282e-04 -3.698872e-01 -3.919506e-03  6.721700e-02  4.378847e-03 -8.939731e-02 -2.829340e+00
  -3.727596e-03  3.828670e-01 -2.649899e-04  1.974183e-01 -1.426923e-03  1.112492e-01  3.462102e-03 -1.197483e+01
  -5.587930e-02 -9.202884e-04 -1.680494e-01 -9.393852e-03 -7.746791e-02  5.855954e-03 -9.722956e-02 -5.848163e+00
   5.552556e-04 -1.777997e-01 -1.259961e-03  5.495508e-02 -6.984099e-03  1.789517e-01  9.278401e-04 -4.013049e+01
  -8.557728e-03 -5.683042e-05 -3.781561e-02  1.895281e-03  1.577835e-01 -3.637657e-03 -1.105895e-02 -5.261157e+00
  -3.201470e-05 -5.429765e-02  1.224458e-03  3.876620e-01  8.853239e-03  2.208250e-01  7.029536e-03 -1.467221e+01
   1.201616e-02 -2.774000e-03 -2.155534e-01  5.910518e-03 -1.760124e-01  7.430681e-04  1.183353e-01 -5.101052e+00
   1.910125e-04 -1.547499e-02 -2.477674e-04  3.447358e-03  2.446402e-04 -3.410851e-04  1.363043e-04 -6.138985e-01
   1.692993e+00  1.338352e-02  5.249003e-02 -1.413601e-02  1.360782e-01  5.228264e-03 -1.091880e-01 -4.476002e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  6427 / 6427 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 2.99e+00       1    0.47
s(measurement.no):toneBis.ord1                      9.00e+00 1.70e+00       1    0.51
s(measurement.no):toneBis.ord2                      9.00e+00 1.00e+00       1    0.53
s(measurement.no):toneBis.ord3                      9.00e+00 1.26e+00       1    0.47
s(measurement.no,speaker)                           1.00e+02 5.50e+01       1    0.46
s(measurement.no,speaker):toneBis.ord1              1.00e+02 6.84e-04       1    0.43
s(measurement.no,speaker):toneBis.ord2              1.00e+02 4.12e+00       1    0.41
s(measurement.no,speaker):toneBis.ord3              1.00e+02 3.83e+01       1    0.46
s(measurement.no,speakerLeftRightTone)              5.90e+02 5.87e+01       1    0.48
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.90e+02 2.26e+01       1    0.53
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.90e+02 3.42e+01       1    0.46
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.90e+02 5.66e+00       1    0.43
s(measurement.no,speakerPos)                        3.00e+02 3.56e+01       1    0.48
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 9.08e+01       1    0.36
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 3.95e+01       1    0.45
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 1.02e+01       1    0.46
s(measurement.no,word)                              6.15e+02 2.60e+02       1    0.45
s(measurement.no,wordPos)                           9.33e+02 1.99e+02       1    0.46
s(measurement.no,wordLeftRightTone)                 8.79e+02 1.90e+02       1    0.42

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1f, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F2 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1f, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1f, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1f, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting 
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1f, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(1700, 2050), xlab = "Time (ms)", ylab = "F2 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2f * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1f, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4f * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1f, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1f * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1f, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3f * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("bottomright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1f, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1f, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1f, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1f, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1f, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1f, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F2 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1f, pred = list(toneBis.ord=c("1", "2", "3", "4")), xlab = "F2 (Z)", main = "Tone")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.8 Model 1G: /au/ male, F2 as output

gamm.model1g.noAR <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1g.noAR, paste("Gamm_model1g_noAR.rds"))

gamm.model1g.noAR <- 
  readRDS("Gamm_model1g_noAR.rds")

r.gamm.model1g <- start_value_rho(gamm.model1g.noAR)

# Auto-regressive model

gamm.model1g <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                    
                      data=data.au.mas, method="fREML", rho = r.gamm.model1g, AR.start = data.au.mas$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1g, paste("Gamm_model1g.rds"))

gamm.model1g <- 
  readRDS("Gamm_model1g.rds")

summary(gamm.model1g, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)  
(Intercept)   0.04924    0.06351   0.775   0.4381  
toneBis.ord1 -0.19064    0.10813  -1.763   0.0779 .
toneBis.ord2 -0.19199    0.16524  -1.162   0.2453  
toneBis.ord3 -0.09237    0.13347  -0.692   0.4889  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df      F p-value    
s(measurement.no)              7.165  7.777 54.388  <2e-16 ***
s(measurement.no):toneBis.ord1 1.616  1.831  0.119   0.854    
s(measurement.no):toneBis.ord2 1.698  1.966  0.519   0.593    
s(measurement.no):toneBis.ord3 2.001  2.377  1.142   0.274    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.726   Deviance explained =   77%
fREML = 9170.2  Scale est. = 0.24868   n = 10857

gam.check(gamm.model1g)

Method: fREML   Optimizer: perf chol
$grad
 [1]  2.229328e-13 -1.039169e-13 -1.226241e-13  4.013456e-13 -5.329071e-15 -2.220446e-14  9.592327e-14  1.310063e-14 -4.903654e-05
[10] -7.888230e-05 -9.703349e-14  7.105427e-14 -3.197442e-13  1.207923e-13 -1.350031e-13  2.486900e-13 -5.989232e-05 -7.860379e-14
[19] -9.947598e-14  3.019807e-14  3.197442e-14 -4.796163e-14 -6.579657e-05 -5.042393e-05  1.030287e-13  4.618528e-14  5.329071e-14
[28]  1.394440e-13 -3.126388e-13 -1.008971e-12  1.421085e-13  4.689582e-13 -5.684342e-14  7.105427e-13  9.094947e-12

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.437874e+00  2.316266e-02  2.018002e-02 -5.515869e-02 -6.066135e-02  4.707938e-06  9.395319e-03  3.562528e-05 -1.505581e-07
   2.316266e-02  3.980832e-02 -1.062052e-02  2.157234e-02 -2.195134e-03 -4.651760e-05  4.614615e-02  1.518743e-04  1.627740e-08
   2.018002e-02 -1.062052e-02  1.623651e-01  4.053160e-02 -1.583566e-02  4.027810e-05 -5.065022e-04 -7.633376e-06 -5.316040e-07
  -5.515869e-02  2.157234e-02  4.053160e-02  2.885601e-01 -4.320236e-03  1.940318e-04  3.109130e-03  5.644838e-05  1.365900e-07
  -6.066135e-02 -2.195134e-03 -1.583566e-02 -4.320236e-03  2.412681e+00  6.689216e-03 -1.198989e-02  9.770154e-04  9.245564e-07
   4.707938e-06 -4.651760e-05  4.027810e-05  1.940318e-04  6.689216e-03  1.311127e+00  1.126068e-04 -1.140562e-01  9.501505e-09
   9.395319e-03  4.614615e-02 -5.065022e-04  3.109130e-03 -1.198989e-02  1.126068e-04  1.795085e+00  5.546141e-03 -4.452620e-07
   3.562528e-05  1.518743e-04 -7.633376e-06  5.644838e-05  9.770154e-04 -1.140562e-01  5.546141e-03  4.739546e-01 -4.754952e-09
  -1.505581e-07  1.627740e-08 -5.316040e-07  1.365900e-07  9.245564e-07  9.501505e-09 -4.452620e-07 -4.754952e-09  4.903668e-05
  -7.305871e-10  7.964168e-10  7.802123e-08  2.524896e-08 -3.195785e-08 -4.947116e-06 -6.925460e-09 -1.458317e-06  2.026651e-11
  -8.451302e-04  2.044229e-04  4.937841e-04  9.478053e-04  2.268942e-03  3.217867e-05  3.799916e-04 -6.702921e-06  1.758484e-08
  -3.168470e-05  8.746443e-06 -1.699352e-05 -2.364217e-04 -8.192527e-05 -4.396729e-02 -5.381943e-05 -2.052291e-03 -2.847634e-09
   6.200292e-03  1.317073e-02 -3.688721e-02 -1.329708e-02  1.121849e+00  4.031927e-03  7.156981e-02 -6.283709e-04  3.211098e-07
  -6.926706e-05 -2.536088e-04  1.979158e-04  3.175109e-04  2.513075e-03  7.862587e-01 -2.660659e-04 -4.769456e-02  2.947175e-09
   1.726737e-02  6.078346e-02 -3.682938e-03  9.445019e-03 -4.419676e-02 -1.133598e-03  2.439917e+00  6.964696e-03 -6.915297e-07
   3.764560e-05  2.628084e-04 -5.372063e-05  2.126977e-04  1.483001e-03 -1.659597e-01  9.683060e-03  5.094348e-01 -6.350002e-10
   3.076055e-08 -1.095645e-07  1.651019e-06  8.091615e-07  1.465533e-06  1.335006e-08 -1.128683e-06 -1.006685e-08  3.523436e-10
  -3.894257e-05  2.687697e-05  1.729336e-04  4.764481e-05  1.094681e-04 -3.718060e-02 -2.018863e-04 -1.553169e-02  1.200411e-07
   4.130878e-04 -1.087641e-03  6.326398e-03 -3.768399e-02 -7.981084e-03  6.366140e-04 -7.120429e-03  1.463290e-04  1.116032e-06
   2.107742e-05 -5.743470e-06  8.947987e-04 -2.717772e-04  1.678862e-03  1.939915e-03  5.401216e-04  1.474212e-02  2.869437e-10
  -4.975967e-02  1.673679e-02 -1.892785e-03 -1.727756e-02  1.242058e+00  1.890168e-03  1.368316e-01 -2.564841e-04  6.535536e-07
  -7.544341e-05  1.151636e-04 -2.561342e-04  1.642415e-04 -3.447386e-04  2.339544e-01 -2.203239e-04 -1.272270e-03  3.426163e-09
   6.368708e-07  2.135991e-06 -9.643205e-08  2.181347e-07  2.550577e-06  7.479623e-09  5.762013e-05  1.217399e-07 -2.562365e-12
   5.573879e-09  1.694732e-08 -2.672746e-10  7.137451e-09  5.257411e-08 -4.216054e-06 -1.881824e-07  1.372942e-05 -3.541337e-13
  -5.524864e-03 -1.595370e-03  1.310796e-02  2.243005e-02  8.161436e-02  9.707901e-04 -3.689276e-02 -2.824487e-04  1.063025e-05
   3.157755e-06  3.499191e-06  7.249782e-04  4.215737e-04  1.203485e-04 -9.100894e-03 -4.298606e-05 -8.136509e-03  8.770457e-08
  -1.806840e-02  4.151006e-03 -1.477690e-03  2.640669e-02  9.901812e-02  1.390322e-03  2.649766e-02  4.105578e-05  1.051161e-06
  -2.661150e-04  1.332405e-04  1.843904e-04 -1.315551e-03  3.202030e-04 -5.662848e-03  3.045579e-04  4.720731e-03  1.542600e-08
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -7.305871e-10 -8.451302e-04 -3.168470e-05  6.200292e-03 -6.926706e-05  1.726737e-02  3.764560e-05  3.076055e-08 -3.894257e-05
   7.964168e-10  2.044229e-04  8.746443e-06  1.317073e-02 -2.536088e-04  6.078346e-02  2.628084e-04 -1.095645e-07  2.687697e-05
   7.802123e-08  4.937841e-04 -1.699352e-05 -3.688721e-02  1.979158e-04 -3.682938e-03 -5.372063e-05  1.651019e-06  1.729336e-04
   2.524896e-08  9.478053e-04 -2.364217e-04 -1.329708e-02  3.175109e-04  9.445019e-03  2.126977e-04  8.091615e-07  4.764481e-05
  -3.195785e-08  2.268942e-03 -8.192527e-05  1.121849e+00  2.513075e-03 -4.419676e-02  1.483001e-03  1.465533e-06  1.094681e-04
  -4.947116e-06  3.217867e-05 -4.396729e-02  4.031927e-03  7.862587e-01 -1.133598e-03 -1.659597e-01  1.335006e-08 -3.718060e-02
  -6.925460e-09  3.799916e-04 -5.381943e-05  7.156981e-02 -2.660659e-04  2.439917e+00  9.683060e-03 -1.128683e-06 -2.018863e-04
  -1.458317e-06 -6.702921e-06 -2.052291e-03 -6.283709e-04 -4.769456e-02  6.964696e-03  5.094348e-01 -1.006685e-08 -1.553169e-02
   2.026651e-11  1.758484e-08 -2.847634e-09  3.211098e-07  2.947175e-09 -6.915297e-07 -6.350002e-10  3.523436e-10  1.200411e-07
   7.888903e-05  1.439876e-09 -1.851590e-06 -2.050451e-08  2.965460e-06 -1.557946e-08 -1.421767e-07  4.716587e-11  4.348406e-05
   1.439876e-09  4.261686e-03  3.264167e-04 -1.194758e-02  3.281816e-05  1.373393e-04  9.778121e-06  8.854692e-08 -1.156317e-06
  -1.851590e-06  3.264167e-04  1.044634e+00  3.366967e-04 -5.496669e-02 -3.100722e-05  1.454865e-02 -6.256216e-09  4.271824e-03
  -2.050451e-08 -1.194758e-02  3.366967e-04  1.226825e+01  7.700868e-02  7.738119e-01  3.013264e-03  8.265483e-07  4.617527e-04
   2.965460e-06  3.281816e-05 -5.496669e-02  7.700868e-02  1.048073e+01  1.038945e-02 -2.970130e-03  1.024402e-08 -1.257342e-01
  -1.557946e-08  1.373393e-04 -3.100722e-05  7.738119e-01  1.038945e-02  1.201541e+01  5.469194e-02 -1.685569e-06 -8.880102e-05
  -1.421767e-07  9.778121e-06  1.454865e-02  3.013264e-03 -2.970130e-03  5.469194e-02  4.042026e+00  4.713961e-09 -1.747234e-02
   4.716587e-11  8.854692e-08 -6.256216e-09  8.265483e-07  1.024402e-08 -1.685569e-06  4.713961e-09  5.989440e-05  6.459432e-07
   4.348406e-05 -1.156317e-06  4.271824e-03  4.617527e-04 -1.257342e-01 -8.880102e-05 -1.747234e-02  6.459432e-07  1.787920e+00
   7.689642e-08  6.150664e-02  2.798424e-03 -3.992483e-01 -1.138000e-03  6.289388e-02  7.489288e-04  8.302777e-07  8.224739e-05
  -8.815555e-06  2.459809e-04  3.223812e-01  3.147687e-03 -3.882249e-01 -9.079899e-04  9.669501e-03  7.707207e-08 -1.794346e-02
  -4.457709e-08 -2.089977e-04 -2.569032e-04  6.422756e-01  3.763017e-03  4.368643e-01  2.627193e-03  4.379773e-06 -3.161759e-04
   1.087427e-06 -1.232699e-05  3.093683e-02  7.003735e-04  3.476545e-01 -2.551545e-03 -1.757162e-01  4.317841e-09 -5.963117e-03
  -7.880084e-13  2.600275e-08 -7.652363e-09 -3.768505e-06  3.490936e-07  1.057619e-04  8.334761e-07 -1.452046e-11 -2.574785e-08
  -3.450176e-11 -9.090443e-11 -7.768342e-08 -7.261994e-08 -8.597461e-06 -3.632611e-07 -7.706134e-06 -1.317324e-12 -7.595148e-07
   1.483455e-06  4.661171e-03 -6.884880e-04 -2.115217e-01 -9.364835e-04 -5.177068e-02 -1.254979e-05  3.862083e-05  1.175347e-02
   4.047443e-05  3.713670e-05 -1.379101e-02 -2.773177e-04 -3.657457e-02 -1.531191e-04 -1.769128e-02  2.499450e-07  1.872743e-01
   5.323889e-08  8.921723e-02  6.708379e-03 -1.844980e-01  1.048615e-03  1.596218e-02 -2.769383e-05  1.444847e-07  1.005949e-03
   3.682382e-06  1.282240e-04  3.301771e-01  1.360682e-03 -4.707971e-02  1.096983e-03  1.790163e-02  2.165260e-08 -4.448259e-04
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   4.130878e-04  2.107742e-05 -4.975967e-02 -7.544341e-05  6.368708e-07  5.573879e-09 -5.524864e-03  3.157755e-06 -1.806840e-02
  -1.087641e-03 -5.743470e-06  1.673679e-02  1.151636e-04  2.135991e-06  1.694732e-08 -1.595370e-03  3.499191e-06  4.151006e-03
   6.326398e-03  8.947987e-04 -1.892785e-03 -2.561342e-04 -9.643205e-08 -2.672746e-10  1.310796e-02  7.249782e-04 -1.477690e-03
  -3.768399e-02 -2.717772e-04 -1.727756e-02  1.642415e-04  2.181347e-07  7.137451e-09  2.243005e-02  4.215737e-04  2.640669e-02
  -7.981084e-03  1.678862e-03  1.242058e+00 -3.447386e-04  2.550577e-06  5.257411e-08  8.161436e-02  1.203485e-04  9.901812e-02
   6.366140e-04  1.939915e-03  1.890168e-03  2.339544e-01  7.479623e-09 -4.216054e-06  9.707901e-04 -9.100894e-03  1.390322e-03
  -7.120429e-03  5.401216e-04  1.368316e-01 -2.203239e-04  5.762013e-05 -1.881824e-07 -3.689276e-02 -4.298606e-05  2.649766e-02
   1.463290e-04  1.474212e-02 -2.564841e-04 -1.272270e-03  1.217399e-07  1.372942e-05 -2.824487e-04 -8.136509e-03  4.105578e-05
   1.116032e-06  2.869437e-10  6.535536e-07  3.426163e-09 -2.562365e-12 -3.541337e-13  1.063025e-05  8.770457e-08  1.051161e-06
   7.689642e-08 -8.815555e-06 -4.457709e-08  1.087427e-06 -7.880084e-13 -3.450176e-11  1.483455e-06  4.047443e-05  5.323889e-08
   6.150664e-02  2.459809e-04 -2.089977e-04 -1.232699e-05  2.600275e-08 -9.090443e-11  4.661171e-03  3.713670e-05  8.921723e-02
   2.798424e-03  3.223812e-01 -2.569032e-04  3.093683e-02 -7.652363e-09 -7.768342e-08 -6.884880e-04 -1.379101e-02  6.708379e-03
  -3.992483e-01  3.147687e-03  6.422756e-01  7.003735e-04 -3.768505e-06 -7.261994e-08 -2.115217e-01 -2.773177e-04 -1.844980e-01
  -1.138000e-03 -3.882249e-01  3.763017e-03  3.476545e-01  3.490936e-07 -8.597461e-06 -9.364835e-04 -3.657457e-02  1.048615e-03
   6.289388e-02 -9.079899e-04  4.368643e-01 -2.551545e-03  1.057619e-04 -3.632611e-07 -5.177068e-02 -1.531191e-04  1.596218e-02
   7.489288e-04  9.669501e-03  2.627193e-03 -1.757162e-01  8.334761e-07 -7.706134e-06 -1.254979e-05 -1.769128e-02 -2.769383e-05
   8.302777e-07  7.707207e-08  4.379773e-06  4.317841e-09 -1.452046e-11 -1.317324e-12  3.862083e-05  2.499450e-07  1.444847e-07
   8.224739e-05 -1.794346e-02 -3.161759e-04 -5.963117e-03 -2.574785e-08 -7.595148e-07  1.175347e-02  1.872743e-01  1.005949e-03
   4.163459e+00  2.143289e-03 -1.517120e-01 -1.395542e-04  9.327295e-07  6.123641e-09  7.325413e-02  1.263834e-03  2.096324e+00
   2.143289e-03  7.773279e+00  7.615210e-04  4.836602e-02  1.673359e-08  1.629840e-06 -1.978241e-04 -6.983178e-02  8.294863e-04
  -1.517120e-01  7.615210e-04  6.381873e+00 -1.706296e-02  5.122834e-05  1.724851e-07  1.997388e-02  3.903511e-04 -1.316587e-01
  -1.395542e-04  4.836602e-02 -1.706296e-02  2.196855e+00 -6.006520e-07 -6.206556e-07 -1.548556e-03 -1.956458e-01 -1.280320e-03
   9.327295e-07  1.673359e-08  5.122834e-05 -6.006520e-07  6.580981e-05 -3.654483e-11  1.677686e-06 -4.933806e-08  7.725429e-07
   6.123641e-09  1.629840e-06  1.724851e-07 -6.206556e-07 -3.654483e-11  5.042602e-05 -4.414987e-08 -1.261128e-06  5.156091e-09
   7.325413e-02 -1.978241e-04  1.997388e-02 -1.548556e-03  1.677686e-06 -4.414987e-08  6.709925e+00  5.329479e-02  8.642176e-02
   1.263834e-03 -6.983178e-02  3.903511e-04 -1.956458e-01 -4.933806e-08 -1.261128e-06  5.329479e-02  2.156669e+00  1.112837e-03
   2.096324e+00  8.294863e-04 -1.316587e-01 -1.280320e-03  7.725429e-07  5.156091e-09  8.642176e-02  1.112837e-03  7.967526e+00
   1.546872e-03  2.084293e-01 -7.516814e-04 -3.319614e-01 -3.599597e-08  8.111764e-07  1.979182e-03  1.328819e-02  5.692828e-03
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -2.661150e-04  1.163646e-02 -2.188569e-04  1.384030e-02  4.125753e-04  2.297420e-02 -1.364179e-04 -3.082698e+00
   1.332405e-04 -2.710885e-02  3.038571e-04 -4.188369e-02 -2.447702e-04 -4.460484e-02  3.996755e-05 -3.077556e-01
   1.843904e-04  5.811125e-02 -2.917499e-04  1.671158e-02 -5.151884e-04 -5.437552e-02  4.881909e-04 -3.489790e-01
  -1.315551e-03  6.089001e-02 -1.374553e-04  3.397011e-02  5.511413e-04  3.513849e-02  2.334519e-03 -5.003965e-01
   3.202030e-04 -1.229868e-02  5.314826e-03 -1.029640e-01  3.437312e-03  4.422363e-02  2.151496e-03 -8.700154e+00
  -5.662848e-03  9.824874e-04 -5.295970e-02  3.702037e-05 -1.907415e-01  2.056159e-03 -1.317421e-01 -2.465283e+00
   3.045579e-04 -2.920378e-02 -2.196896e-03  6.687497e-02 -1.784388e-03 -1.741749e-01 -2.235248e-04 -7.382928e+00
   4.720731e-03  1.524239e-03 -7.464410e-02  9.422477e-04 -6.590795e-02  2.669401e-03  1.385551e-02 -1.308684e+00
   1.542600e-08  8.663802e-07  4.574252e-08 -2.567976e-06 -1.431636e-08  2.050524e-07  3.747505e-08 -8.270227e-05
   3.682382e-06  1.176618e-07 -5.556982e-06  2.463770e-07 -1.972180e-05 -9.632058e-08 -1.044966e-05 -1.852579e-04
   1.282240e-04  5.902653e-03 -6.322036e-05  1.896552e-02 -7.254751e-05  8.093250e-03  1.067883e-05 -3.279158e-01
   3.301771e-01  2.664065e-04 -3.265108e-01 -2.530802e-03  6.964577e-02  3.757299e-03  1.256601e-01 -2.124746e+00
   1.360682e-03  4.006361e-01  2.401440e-03  4.435018e-01  3.171134e-03  1.718547e+00 -4.395128e-04 -4.029830e+01
  -4.707971e-02 -1.119001e-03 -9.524887e-01  6.070144e-04 -2.621125e-01  2.795584e-02 -2.727400e-01 -1.621998e+01
   1.096983e-03  4.843734e-01  1.277907e-04  5.999173e-01  9.479997e-03  7.503544e-01  1.318917e-03 -3.813422e+01
   1.790163e-02  1.615325e-03  1.317023e-01  4.936278e-03 -1.653512e-01  1.717751e-02  2.224564e-01 -8.852333e+00
   2.165260e-08  2.156973e-05  1.531343e-07  1.216068e-05  3.237014e-08  2.347271e-05  2.333424e-07 -3.553942e-04
  -4.448259e-04  4.596092e-03  3.892882e-01  7.690336e-03  2.065172e-01  9.789057e-03  2.340366e-01 -3.969899e+00
   1.546872e-03  3.602447e-01  1.319812e-04  3.877252e-01 -8.422117e-04  2.043456e+00  2.542471e-04 -1.895627e+01
   2.084293e-01  1.145925e-03 -1.516239e+00 -2.342256e-02  2.249494e+00  1.956533e-02  4.084740e-01 -1.378275e+01
  -7.516814e-04  5.961189e-03  6.372762e-03 -1.804802e-01  3.859804e-03  1.178360e-01  2.061606e-03 -2.150542e+01
  -3.319614e-01  8.753924e-04 -5.096577e-01  8.267062e-03 -1.487580e-01 -3.576956e-03 -5.207596e-02 -5.046148e+00
  -3.599597e-08 -4.041343e-08  4.452367e-08  4.030344e-05  4.795860e-07 -1.185712e-05 -7.455456e-07 -8.628078e-04
   8.111764e-07  1.659950e-07  5.315215e-06  5.092938e-08  1.455300e-05  8.230839e-07  2.047104e-05 -1.994034e-04
   1.979182e-03  1.403369e-01  5.380428e-03 -2.327774e-01  1.238792e-02  2.722891e-01  5.228962e-03 -2.433305e+01
   1.328819e-02  2.394346e-03  4.751751e-02  6.237748e-03  2.248064e-01  4.839937e-03  1.971381e-01 -4.246845e+00
   5.692828e-03  1.442915e-02 -1.070577e-03 -1.127369e-01 -4.036508e-03  6.703618e-03 -3.794700e-03 -2.351528e+01
   4.696572e+00  1.986515e-03 -7.995356e-01 -6.053670e-03 -3.215859e-01 -7.179484e-03 -1.594099e-02 -7.194074e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  7802 / 7802 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 7.17e+00    1.03    0.96
s(measurement.no):toneBis.ord1                      9.00e+00 1.62e+00    1.03    0.98
s(measurement.no):toneBis.ord2                      9.00e+00 1.70e+00    1.03    0.97
s(measurement.no):toneBis.ord3                      9.00e+00 2.00e+00    1.03    0.99
s(measurement.no,speaker)                           1.00e+02 2.23e+01    1.03    0.98
s(measurement.no,speaker):toneBis.ord1              1.00e+02 1.74e+01    1.03    0.98
s(measurement.no,speaker):toneBis.ord2              1.00e+02 7.92e-04    1.03    0.97
s(measurement.no,speaker):toneBis.ord3              1.00e+02 4.91e+00    1.03    0.99
s(measurement.no,speakerLeftRightTone)              6.00e+02 1.13e+02    1.03    0.98
s(measurement.no,speakerLeftRightTone):toneBis.ord1 6.00e+02 9.40e+01    1.03    0.97
s(measurement.no,speakerLeftRightTone):toneBis.ord2 6.00e+02 7.94e+00    1.03    0.97
s(measurement.no,speakerLeftRightTone):toneBis.ord3 6.00e+02 6.55e+01    1.03    0.99
s(measurement.no,speakerPos)                        3.00e+02 5.31e+01    1.03    0.98
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 2.36e-03    1.03    0.99
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 5.72e+01    1.03    0.99
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 6.14e+01    1.03    0.96
s(measurement.no,word)                              9.75e+02 3.38e+02    1.03    0.98
s(measurement.no,wordPos)                           1.43e+03 4.17e+02    1.03    0.95
s(measurement.no,wordLeftRightTone)                 1.36e+03 5.01e+02    1.03    0.99

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1g, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F2 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1g, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1g, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1g, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1g, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(1050, 1350), xlab = "Time (ms)", ylab = "F2 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1au * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1g, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4au * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1g, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2au * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1g, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3au * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("top", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1g, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1g, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1g, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1g, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1g, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

plot_diff(gamm.model1g, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F2 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

NA
NA

plot_parametric(gamm.model1g, pred = list(toneBis.ord=c("1", "2", "3", "4")), xlab = "F2 (Z)", main = "Tone")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

4.9 Model 1H: /au/ female, F2 as output

gamm.model1h <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                        
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1h.noAR, paste("Gamm_model1h_noAR.rds"))

gamm.model1h.noAR <- 
  readRDS("Gamm_model1h_noAR.rds")

r.gamm.model1h <- start_value_rho(gamm.model1h.noAR)

# Auto-regressive model

gamm.model1h <- bam(f2Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr") +
                        
                        # random effects
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                        
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +  
                        
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1, by=toneBis.ord) +
                          
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                          
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1),
                    
                      data=data.au.fem, method="fREML", rho = r.gamm.model1h, AR.start = data.au.fem$start, discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model1h, paste("Gamm_model1h.rds"))

gamm.model1h <- 
  readRDS("Gamm_model1h.rds")

summary(gamm.model1h, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ toneBis.ord + s(measurement.no, bs = "cr") + s(measurement.no, 
    by = toneBis.ord, bs = "cr") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1, by = toneBis.ord) + s(measurement.no, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1, 
    by = toneBis.ord) + s(measurement.no, word, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.05552    0.04806   1.155 0.248053    
toneBis.ord1 -0.31831    0.08609  -3.697 0.000219 ***
toneBis.ord2 -0.16235    0.13889  -1.169 0.242453    
toneBis.ord3  0.02006    0.08500   0.236 0.813438    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                 edf Ref.df      F p-value    
s(measurement.no)              7.332  7.724 32.386 < 2e-16 ***
s(measurement.no):toneBis.ord1 3.735  4.515  1.410 0.25469    
s(measurement.no):toneBis.ord2 1.000  1.000  8.169 0.00427 ** 
s(measurement.no):toneBis.ord3 3.550  4.198  3.324 0.00813 ** 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.744   Deviance explained = 78.7%
fREML =   8049  Scale est. = 0.2311    n = 10252

gam.check(gamm.model1h)

Method: fREML   Optimizer: perf chol
$grad
 [1] -2.056133e-13 -9.925394e-14 -5.381781e-05  7.482903e-14  2.238210e-13 -9.080208e-05 -5.604331e-05 -1.954764e-06 -3.857644e-05
[10] -5.551115e-15  3.552714e-15 -7.882583e-15  2.771117e-13  1.776357e-14 -4.618528e-13 -3.019807e-14 -1.953993e-13 -9.503509e-14
[19] -3.552714e-14 -8.881784e-15 -2.202682e-13 -6.661338e-14 -7.620288e-05 -2.547962e-14 -5.329071e-14 -9.316737e-05 -1.172396e-13
[28] -2.753353e-14  5.684342e-14 -1.989520e-13 -6.536993e-13 -1.136868e-13 -1.705303e-13  7.105427e-15 -5.456968e-12

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   2.886506e+00 -8.165956e-02  4.205714e-06  4.235402e-02  3.969066e-02  7.414044e-09 -9.806132e-07 -5.454388e-09  9.803514e-07
  -8.165956e-02  5.854181e-01  3.361882e-06  4.546892e-02 -1.799790e-02 -1.986621e-08  5.026616e-06 -6.151380e-08  5.854191e-07
   4.205714e-06  3.361882e-06  5.381423e-05 -8.825209e-06 -2.121406e-06 -5.018114e-12 -2.242107e-11  1.066201e-12  2.452243e-10
   4.235402e-02  4.546892e-02 -8.825209e-06  9.564489e-01 -2.190838e-02 -3.053887e-08  3.316095e-07 -5.468488e-08 -1.097971e-06
   3.969066e-02 -1.799790e-02 -2.121406e-06 -2.190838e-02  1.300605e+01 -1.472137e-06 -9.948271e-07 -8.124166e-07 -3.023621e-06
   7.414044e-09 -1.986621e-08 -5.018114e-12 -3.053887e-08 -1.472137e-06  9.080134e-05 -2.037248e-12 -7.149893e-09  1.366655e-12
  -9.806132e-07  5.026616e-06 -2.242107e-11  3.316095e-07 -9.948271e-07 -2.037248e-12  5.604629e-05  6.602073e-10 -3.924320e-11
  -5.454388e-09 -6.151380e-08  1.066201e-12 -5.468488e-08 -8.124166e-07 -7.149893e-09  6.602073e-10  2.223720e-06 -4.942095e-12
   9.803514e-07  5.854191e-07  2.452243e-10 -1.097971e-06 -3.023621e-06  1.366655e-12 -3.924320e-11 -4.942095e-12  3.858004e-05
   1.396230e-05 -1.406249e-05  6.633962e-08 -1.900635e-04 -7.678157e-04 -3.554136e-07  4.994476e-09  2.585219e-06  7.533450e-08
   9.376440e-03  3.573732e-03 -6.001278e-07  2.180237e-02  2.655765e-02  7.471867e-08  8.484395e-07  2.837683e-07 -7.046646e-07
  -6.808330e-06 -1.959267e-05  7.773505e-09  1.268144e-04 -2.564425e-04 -1.033468e-06  2.932959e-10  1.470165e-06  1.861346e-09
   1.072305e-01 -2.293115e-03 -1.249674e-06  2.594501e-02  2.915698e+00 -4.124312e-07 -2.296648e-06 -2.553101e-07 -1.367543e-05
  -5.113264e-05  1.407438e-04 -4.736159e-09 -4.440147e-04 -2.015467e-02  7.257574e-06 -8.948578e-08 -4.725988e-05  7.320647e-08
  -1.206730e-02  1.124403e-01 -1.493274e-06  1.918102e-04  1.506926e-01 -1.993904e-07  6.519124e-05  1.284654e-05  5.939598e-07
   1.142488e-04  5.683852e-04  6.149664e-09  2.246579e-04 -3.118093e-03 -1.130600e-05  1.477604e-06  5.663082e-04  8.979453e-09
   2.200695e-02 -1.333230e-02  1.241474e-05 -1.086276e-02 -7.985877e-03  1.077522e-07 -1.042582e-06 -3.797280e-07  1.185975e-04
   3.913438e-05  1.639885e-06  1.393471e-07 -4.512330e-04 -2.294813e-03 -2.518078e-06  1.972913e-08  2.624094e-06  3.919648e-07
   1.777503e-02 -4.356595e-03 -2.078651e-06  1.281784e-01 -1.006199e-02  2.349787e-07  6.891805e-09  3.842853e-07 -1.808104e-06
  -3.862066e-04 -7.429886e-04  1.297983e-08  2.257717e-03  1.347356e-03 -6.810254e-06 -1.810731e-08  2.214223e-05  1.053697e-08
  -1.469443e-02  2.264674e-03  2.490395e-06  1.501686e-02  1.841689e+00 -2.522577e-07  4.834629e-06 -1.025341e-06  4.169355e-06
   2.984468e-04  4.180668e-05 -2.001799e-07 -6.658631e-04 -9.969294e-03 -7.344100e-06 -8.820564e-08 -3.436731e-05  1.314116e-08
  -1.070262e-06  3.650026e-06 -2.750601e-11  5.192212e-07  1.308526e-05  9.138350e-13  3.188189e-09  7.804734e-10 -1.924734e-11
   5.899755e-06 -4.036605e-05 -7.390840e-10 -1.552603e-06  3.510216e-04 -5.981947e-07  8.665276e-08  2.355309e-05  3.739131e-10
   9.899997e-03  4.437641e-03  8.036792e-06 -6.254622e-03 -2.712646e-02 -4.790932e-08 -7.759347e-07 -1.881713e-07  6.061738e-05
  -3.004297e-08 -8.786602e-08  1.636921e-11  2.189748e-07 -5.196137e-08  7.865171e-11  4.153039e-12  1.002061e-09  2.358607e-11
   1.748174e-03  2.190495e-04 -1.256478e-07  2.572225e-02  3.933658e-02  5.137537e-09  1.373502e-07  1.381352e-07 -7.890231e-07
  -1.017652e-04 -4.262670e-04  6.334965e-08  1.175215e-03 -1.018894e-03 -3.479467e-06 -1.500264e-08  7.517212e-06  2.536061e-08
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   1.396230e-05  9.376440e-03 -6.808330e-06  1.072305e-01 -5.113264e-05 -1.206730e-02  1.142488e-04  2.200695e-02  3.913438e-05
  -1.406249e-05  3.573732e-03 -1.959267e-05 -2.293115e-03  1.407438e-04  1.124403e-01  5.683852e-04 -1.333230e-02  1.639885e-06
   6.633962e-08 -6.001278e-07  7.773505e-09 -1.249674e-06 -4.736159e-09 -1.493274e-06  6.149664e-09  1.241474e-05  1.393471e-07
  -1.900635e-04  2.180237e-02  1.268144e-04  2.594501e-02 -4.440147e-04  1.918102e-04  2.246579e-04 -1.086276e-02 -4.512330e-04
  -7.678157e-04  2.655765e-02 -2.564425e-04  2.915698e+00 -2.015467e-02  1.506926e-01 -3.118093e-03 -7.985877e-03 -2.294813e-03
  -3.554136e-07  7.471867e-08 -1.033468e-06 -4.124312e-07  7.257574e-06 -1.993904e-07 -1.130600e-05  1.077522e-07 -2.518078e-06
   4.994476e-09  8.484395e-07  2.932959e-10 -2.296648e-06 -8.948578e-08  6.519124e-05  1.477604e-06 -1.042582e-06  1.972913e-08
   2.585219e-06  2.837683e-07  1.470165e-06 -2.553101e-07 -4.725988e-05  1.284654e-05  5.663082e-04 -3.797280e-07  2.624094e-06
   7.533450e-08 -7.046646e-07  1.861346e-09 -1.367543e-05  7.320647e-08  5.939598e-07  8.979453e-09  1.185975e-04  3.919648e-07
   7.933819e-02 -2.047914e-04  1.096834e-03 -1.195762e-04 -1.883057e-02  1.730831e-04  7.190810e-03  4.068010e-03  4.107291e-01
  -2.047914e-04  5.351787e-01  2.594692e-03  1.282930e-02  1.684367e-04  1.461684e-02 -1.514458e-04 -4.238278e-03 -5.753731e-05
   1.096834e-03  2.594692e-03  2.048069e-02  3.334529e-05 -1.156569e-02 -5.072850e-05  2.091834e-03  2.170208e-06  8.626483e-03
  -1.195762e-04  1.282930e-02  3.334529e-05  1.435264e+01  4.791918e-02  1.007960e+00  1.982998e-03 -6.172665e-01 -3.907466e-03
  -1.883057e-02  1.684367e-04 -1.156569e-02  4.791918e-02  8.618028e+00  8.347341e-06 -7.081581e-04  4.346679e-03 -1.938123e-01
   1.730831e-04  1.461684e-02 -5.072850e-05  1.007960e+00  8.347341e-06  1.162960e+01  5.817242e-02 -2.117739e-02  1.616569e-03
   7.190810e-03 -1.514458e-04  2.091834e-03  1.982998e-03 -7.081581e-04  5.817242e-02  5.746728e+00  5.947032e-04  3.575333e-02
   4.068010e-03 -4.238278e-03  2.170208e-06 -6.172665e-01  4.346679e-03 -2.117739e-02  5.947032e-04  7.532931e+00  5.418699e-02
   4.107291e-01 -5.753731e-05  8.626483e-03 -3.907466e-03 -1.938123e-01  1.616569e-03  3.575333e-02  5.418699e-02  3.116018e+00
   9.411679e-04  9.884525e-01  3.781443e-03 -3.380966e-01  1.804735e-03  5.765664e-02 -8.456714e-05  1.134806e-01  7.192935e-03
   3.142429e-02  1.632188e-02  1.083723e-01 -2.615276e-03 -6.221225e-01 -3.002136e-03  5.590658e-02  2.397540e-03  1.606300e-01
   2.711764e-04  4.154280e-02  9.510881e-05  1.007055e+00 -1.374389e-02 -2.103386e-01 -6.217315e-03  1.506407e-01  3.570076e-03
   2.714099e-02  1.105886e-03 -8.834764e-03  2.218211e-03  2.703374e-01 -5.023120e-03 -1.831192e-01  2.369937e-03  1.312044e-01
  -3.613284e-09  8.300473e-07 -1.179357e-09  1.146377e-05 -5.910111e-08  1.075408e-04  1.586930e-06 -1.131739e-07 -3.705227e-08
   1.661186e-03  3.244934e-07  8.194045e-05  1.092027e-04 -1.394831e-02  1.660074e-03  7.754395e-02  6.928677e-05  9.060316e-03
   2.449292e-03  2.168689e-02  2.649863e-04 -1.497470e-01  1.881975e-03 -4.552947e-03  9.060298e-04  2.714276e+00  2.794188e-02
   6.171795e-05  1.465217e-06 -3.277593e-06 -8.284783e-07  1.877677e-05  4.280971e-07 -3.755343e-07  6.161042e-06  4.014409e-04
   1.349794e-04  2.393237e-01  1.184016e-03  1.494030e-01  3.073656e-04 -6.045488e-06  1.858074e-04 -8.130998e-02  1.071648e-03
   4.450671e-03  1.235016e-02  9.063413e-02 -2.364311e-03 -8.753786e-02 -1.215969e-03 -9.845598e-02  5.441616e-04  1.194015e-02
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   1.777503e-02 -3.862066e-04 -1.469443e-02  2.984468e-04 -1.070262e-06  5.899755e-06  9.899997e-03 -3.004297e-08  1.748174e-03
  -4.356595e-03 -7.429886e-04  2.264674e-03  4.180668e-05  3.650026e-06 -4.036605e-05  4.437641e-03 -8.786602e-08  2.190495e-04
  -2.078651e-06  1.297983e-08  2.490395e-06 -2.001799e-07 -2.750601e-11 -7.390840e-10  8.036792e-06  1.636921e-11 -1.256478e-07
   1.281784e-01  2.257717e-03  1.501686e-02 -6.658631e-04  5.192212e-07 -1.552603e-06 -6.254622e-03  2.189748e-07  2.572225e-02
  -1.006199e-02  1.347356e-03  1.841689e+00 -9.969294e-03  1.308526e-05  3.510216e-04 -2.712646e-02 -5.196137e-08  3.933658e-02
   2.349787e-07 -6.810254e-06 -2.522577e-07 -7.344100e-06  9.138350e-13 -5.981947e-07 -4.790932e-08  7.865171e-11  5.137537e-09
   6.891805e-09 -1.810731e-08  4.834629e-06 -8.820564e-08  3.188189e-09  8.665276e-08 -7.759347e-07  4.153039e-12  1.373502e-07
   3.842853e-07  2.214223e-05 -1.025341e-06 -3.436731e-05  7.804734e-10  2.355309e-05 -1.881713e-07  1.002061e-09  1.381352e-07
  -1.808104e-06  1.053697e-08  4.169355e-06  1.314116e-08 -1.924734e-11  3.739131e-10  6.061738e-05  2.358607e-11 -7.890231e-07
   9.411679e-04  3.142429e-02  2.711764e-04  2.714099e-02 -3.613284e-09  1.661186e-03  2.449292e-03  6.171795e-05  1.349794e-04
   9.884525e-01  1.632188e-02  4.154280e-02  1.105886e-03  8.300473e-07  3.244934e-07  2.168689e-02  1.465217e-06  2.393237e-01
   3.781443e-03  1.083723e-01  9.510881e-05 -8.834764e-03 -1.179357e-09  8.194045e-05  2.649863e-04 -3.277593e-06  1.184016e-03
  -3.380966e-01 -2.615276e-03  1.007055e+00  2.218211e-03  1.146377e-05  1.092027e-04 -1.497470e-01 -8.284783e-07  1.494030e-01
   1.804735e-03 -6.221225e-01 -1.374389e-02  2.703374e-01 -5.910111e-08 -1.394831e-02  1.881975e-03  1.877677e-05  3.073656e-04
   5.765664e-02 -3.002136e-03 -2.103386e-01 -5.023120e-03  1.075408e-04  1.660074e-03 -4.552947e-03  4.280971e-07 -6.045488e-06
  -8.456714e-05  5.590658e-02 -6.217315e-03 -1.831192e-01  1.586930e-06  7.754395e-02  9.060298e-04 -3.755343e-07  1.858074e-04
   1.134806e-01  2.397540e-03  1.506407e-01  2.369937e-03 -1.131739e-07  6.928677e-05  2.714276e+00  6.161042e-06 -8.130998e-02
   7.192935e-03  1.606300e-01  3.570076e-03  1.312044e-01 -3.705227e-08  9.060316e-03  2.794188e-02  4.014409e-04  1.071648e-03
   1.450465e+01  8.068396e-02 -2.322966e-01 -1.120044e-04  1.546342e-06 -1.326965e-04  1.352447e-01  2.191644e-06  1.472407e+00
   8.068396e-02  4.994323e+00 -1.354518e-03 -7.528376e-02 -1.449636e-07 -2.281753e-04  1.081392e-02  2.796713e-05  3.716360e-03
  -2.322966e-01 -1.354518e-03  6.311074e+00  1.162652e-02 -2.302263e-05 -9.332746e-04  2.509813e-01  1.159615e-06 -3.984301e-01
  -1.120044e-04 -7.528376e-02  1.162652e-02  3.613317e+00 -1.749931e-07 -5.564151e-02 -6.419589e-03  3.759500e-05  1.833572e-03
   1.546342e-06 -1.449636e-07 -2.302263e-05 -1.749931e-07  7.621419e-05  2.528416e-08  7.251031e-08 -9.692229e-12  1.152698e-06
  -1.326965e-04 -2.281753e-04 -9.332746e-04 -5.564151e-02  2.528416e-08  2.120648e-02  1.300634e-04  1.346675e-06 -3.780534e-05
   1.352447e-01  1.081392e-02  2.509813e-01 -6.419589e-03  7.251031e-08  1.300634e-04  2.841610e+00  1.035599e-06  9.122822e-03
   2.191644e-06  2.796713e-05  1.159615e-06  3.759500e-05 -9.692229e-12  1.346675e-06  1.035599e-06  9.327441e-05  1.760766e-06
   1.472407e+00  3.716360e-03 -3.984301e-01  1.833572e-03  1.152698e-06 -3.780534e-05  9.122822e-03  1.760766e-06  1.196596e+00
   1.235850e-02  7.557761e-02 -2.835687e-03  2.676195e-01  4.582238e-08 -1.108945e-02  1.217038e-02  4.661362e-05  1.383148e-02
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]
  -1.017652e-04  4.258808e-02 -1.360150e-03  3.322268e-03 -9.971456e-06  7.513848e-03 -6.354940e-04 -3.165894e+00
  -4.262670e-04  9.882143e-02 -1.016725e-03  3.402620e-02  6.864391e-04  8.031297e-02 -1.222000e-03 -1.367253e+00
   6.334965e-08  4.306426e-07 -1.592972e-07 -9.994174e-06 -5.212644e-08 -1.112493e-05  2.342457e-07 -6.390932e-05
   1.175215e-03 -1.533348e-01  1.645684e-03 -1.380154e-02 -6.476801e-04 -1.070087e-01  5.018938e-04 -1.275058e+00
  -1.018894e-03  9.345435e-02  6.575828e-03 -2.739587e-01  1.417134e-03  1.539258e-01  3.241403e-03 -1.976019e+01
  -3.479467e-06  8.930580e-08 -1.212004e-05  3.019437e-07  4.748293e-06  1.667192e-08  5.186820e-06 -8.710142e-05
  -1.500264e-08  1.796511e-06 -1.500460e-08  1.140299e-05 -5.597067e-08 -3.728870e-06  5.989330e-09 -3.106739e-04
   7.517212e-06  2.448654e-06 -5.116654e-05  2.348723e-07 -3.071005e-05  2.768391e-06  7.628962e-05 -1.035268e-03
   2.536061e-08 -4.790622e-06  1.103795e-07 -5.536293e-06 -7.359308e-09 -1.054108e-05 -6.506780e-08 -3.032537e-04
   4.450671e-03 -5.112386e-04 -1.321889e-02  5.654700e-04 -2.250108e-02 -2.140277e-04  3.746993e-02 -7.558734e-01
   1.235016e-02  2.054230e-02  1.758034e-03 -1.710042e-02 -1.830262e-03  9.673513e-02 -9.864212e-04 -4.159457e+00
   9.063413e-02  2.799252e-04 -2.023547e-03 -6.154558e-05  1.481228e-02  4.067976e-04  9.443481e-03 -3.076511e-01
  -2.364311e-03  1.414336e+00 -1.346433e-02  4.423084e-01 -2.125935e-03  3.110768e+00  6.736847e-03 -4.865230e+01
  -8.753786e-02  3.486488e-03 -6.946701e-01  1.001795e-02 -3.000997e-01  2.918616e-02 -2.431340e-01 -1.410859e+01
  -1.215969e-03  3.614122e-01  6.535554e-03  9.034376e-01 -3.426288e-03  5.183498e-01  3.993955e-03 -4.683977e+01
  -9.845598e-02  2.359391e-02  3.703734e-01  9.277112e-03 -1.266902e-01  3.311281e-02  6.289435e-01 -1.136494e+01
   5.441616e-04  7.550260e-01  1.050738e-02  2.482924e-01 -2.646575e-04  1.089313e+00  2.023632e-03 -2.187818e+01
   1.194015e-02  9.607222e-04  7.693279e-02  6.302139e-03 -4.182787e-02  1.570391e-02  4.390413e-01 -5.852202e+00
   1.235850e-02  6.642941e-01 -1.679143e-03  6.888582e-01  1.577090e-03  9.456626e-02  8.396634e-03 -4.769336e+01
   7.557761e-02  6.317578e-03 -1.998857e-01  1.910147e-02  3.861191e-01  1.644910e-02 -2.603969e-02 -1.099547e+01
  -2.835687e-03  1.883312e-01  5.845969e-03  3.252394e-01  4.116890e-03  8.603551e-01  6.514449e-03 -2.017136e+01
   2.676195e-01  7.188005e-04 -1.599402e-01 -7.030199e-04 -2.095657e-01  3.853118e-03 -1.421315e-01 -6.970679e+00
   4.582238e-08  1.889327e-05  2.476178e-07  5.242289e-05  9.862300e-10 -9.372364e-06 -1.135426e-08 -8.807661e-04
  -1.108945e-02 -6.490300e-04 -5.122356e-02  1.678026e-03  1.880673e-02 -1.564464e-03 -6.453308e-04 -4.583916e-01
   1.217038e-02  1.059145e-01  1.631450e-02  3.379487e-02 -1.028514e-02  9.566842e-02 -4.237459e-03 -1.303378e+01
   4.661362e-05  1.130094e-06  2.785413e-04 -3.245905e-06 -1.763910e-05 -2.415180e-06  2.538859e-05 -1.262042e-03
   1.383148e-02 -3.375729e-02  2.344395e-03  4.313759e-01  1.717118e-03  1.301346e-01  1.109539e-03 -9.207264e+00
   3.299539e+00 -3.034153e-03 -5.597238e-01  8.277296e-03  1.232575e-01 -5.090926e-04 -1.322615e-01 -6.145576e+00
 [ getOption("max.print") est atteint -- 7 lignes omises ]

Model rank =  7542 / 7542 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                          k'      edf k-index p-value
s(measurement.no)                                   9.00e+00 7.33e+00    0.99    0.17
s(measurement.no):toneBis.ord1                      9.00e+00 3.73e+00    0.99    0.17
s(measurement.no):toneBis.ord2                      9.00e+00 1.00e+00    0.99    0.10
s(measurement.no):toneBis.ord3                      9.00e+00 3.55e+00    0.99    0.16
s(measurement.no,speaker)                           1.00e+02 3.95e+01    0.99    0.17
s(measurement.no,speaker):toneBis.ord1              1.00e+02 2.81e-03    0.99    0.21
s(measurement.no,speaker):toneBis.ord2              1.00e+02 1.51e+00    0.99    0.17
s(measurement.no,speaker):toneBis.ord3              1.00e+02 8.93e+00    0.99    0.18
s(measurement.no,speakerLeftRightTone)              5.80e+02 1.26e+02    0.99    0.14
s(measurement.no,speakerLeftRightTone):toneBis.ord1 5.80e+02 1.16e+02    0.99    0.15
s(measurement.no,speakerLeftRightTone):toneBis.ord2 5.80e+02 5.55e+01    0.99    0.19
s(measurement.no,speakerLeftRightTone):toneBis.ord3 5.80e+02 1.17e+02    0.99    0.13
s(measurement.no,speakerPos)                        3.00e+02 5.43e+01    0.99    0.15
s(measurement.no,speakerPos):toneBis.ord1           3.00e+02 9.19e-01    0.99    0.12
s(measurement.no,speakerPos):toneBis.ord2           3.00e+02 2.61e+01    0.99    0.21
s(measurement.no,speakerPos):toneBis.ord3           3.00e+02 3.07e+01    0.99    0.10
s(measurement.no,word)                              9.30e+02 3.86e+02    0.99    0.19
s(measurement.no,wordPos)                           1.39e+03 3.36e+02    0.99    0.19
s(measurement.no,wordLeftRightTone)                 1.27e+03 4.18e+02    0.99    0.21

# Plotting
# Normalized scale

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1h, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = c(-1, 1), lwd = 4, xlab = "Time (normalized)", ylab = "F2 (Z)", xaxt = "n", font.lab = 2)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1h, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1h, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

plot_smooth(gamm.model1h, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

legend("topright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

# Plotting
# Reconstructed scale
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model1h, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), col = "orange", ylim = c(1150, 1600), xlab = "Time (ms)", ylab = "F2 (Hz)", xaxt = "n", font.lab = 2, add = F, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT2auf * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, hide.label = TRUE)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1h, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), col = "royalblue4", lwd = 4, add = T, transform.view = function(measurement.no) measurement.no * durationT4auf * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1h, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT1auf * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model1h, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * durationT3auf * 0.1, transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 
    * Note: X-values are transformed.

legend("top", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=4)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# Plotting
# Difference plot between tones
par(mfcol = c(2, 3), mar = c(2, 2, 2, 1), oma = c(4, 4, 2, 1)) 
plot_diff(gamm.model1h, view="measurement.no", comp=list(toneBis.ord = c("1","2")),rm.ranef=TRUE, main = "1-2", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 3.232323

plot_diff(gamm.model1h, view="measurement.no",comp=list(toneBis.ord = c("1","3")),rm.ranef=TRUE, main = "1-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 4.545455

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
plot_diff(gamm.model1h, view="measurement.no",comp=list(toneBis.ord = c("1","4")),rm.ranef=TRUE, main = "1-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 10.000000

plot_diff(gamm.model1h, view="measurement.no",comp=list(toneBis.ord = c("2","3")),rm.ranef=TRUE, main = "2-3", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    9.191919 - 10.000000

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_diff(gamm.model1h, view="measurement.no",comp=list(toneBis.ord = c("2","4")),rm.ranef=TRUE, main = "2-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    6.666667 - 10.000000

plot_diff(gamm.model1h, view="measurement.no",comp=list(toneBis.ord = c("3","4")),rm.ranef=TRUE, main = "3-4", hide.label = T, xaxt = "n", xlab = "", ylab = "")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

measurement.no window(s) of significant difference(s):
    0.000000 - 1.515152

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
mtext("Difference in F2 (Z)", side = 2, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

mtext("Time (normalized)", side = 1, outer = TRUE, line = 2.5, cex = 1.2, font = 2)

plot_parametric(gamm.model1h, pred = list(toneBis.ord=c("1", "2", "3", "4")), xlab = "F2 (Z)", main = "Tone")

Summary:
    * toneBis.ord : factor; set to the value(s): 1, 2, 3, 4. 
    * measurement.no : numeric predictor; set to the value(s): 5. 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(measurement.no,speaker),s(measurement.no,speaker):toneBis.ord1,s(measurement.no,speaker):toneBis.ord2,s(measurement.no,speaker):toneBis.ord3,s(measurement.no,speakerLeftRightTone),s(measurement.no,speakerLeftRightTone):toneBis.ord1,s(measurement.no,speakerLeftRightTone):toneBis.ord2,s(measurement.no,speakerLeftRightTone):toneBis.ord3,s(measurement.no,speakerPos),s(measurement.no,speakerPos):toneBis.ord1,s(measurement.no,speakerPos):toneBis.ord2,s(measurement.no,speakerPos):toneBis.ord3,s(measurement.no,word),s(measurement.no,wordPos),s(measurement.no,wordLeftRightTone)
 

5 Model 2: Tonal impact on vowel realization from a fine-grained perspective

5.1 Test for Model 2A: Random effects similar to Model 1 (/ai/ male, F1 as output)

gamm.model2test.noAR <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr") +
                        s(f0Zscore2, bs="cr") +
                        s(durationZscore2, bs="cr") +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr") +
                        ti(measurement.no, durationZscore2, bs = "cr") +
                        ti(measurement.no, f0Zscore2, bs = "cr") +
                        ti(f0Zscore2, durationZscore2, bs = "cr") +
                        
                        # random effect
                        
                        ## left segment * pow  
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
 
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                            
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                  
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1),
 
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)

#saveRDS(gamm.model2test.noAR, paste("Gamm_model2test_noAR.rds"))

gamm.model2test.noAR <- 
  readRDS("Gamm_model2test_noAR.rds")

r.gamm.model2test <- start_value_rho(gamm.model2test.noAR)

# Auto-regressive model

gamm.model2test <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr") +
                        s(f0Zscore2, bs="cr") +
                        s(durationZscore2, bs="cr") +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr") +
                        ti(measurement.no, durationZscore2, bs = "cr") +
                        ti(measurement.no, f0Zscore2, bs = "cr") +
                        ti(f0Zscore2, durationZscore2, bs = "cr") +
                        
                        # random effect
                        
                        ## left segment * pow  
                        s(measurement.no, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, word, bs="fs", xt = list(bs="tp"), k=3, m=1) +
 
                        s(measurement.no, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, wordLeftRightTone, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                            
                        s(measurement.no, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(durationZscore2, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                        s(f0Zscore2, wordPos, bs="fs", xt = list(bs="tp"), k=3, m=1) +
                  
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=5, m=1),
                      
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2test, AR.start = data.ai.mas$start)

#saveRDS(gamm.model2test, paste("Gamm_model2test.rds"))

gamm.model2test <- 
  readRDS("Gamm_model2test.rds")

summary(gamm.model2test, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ s(measurement.no, bs = "cr") + s(f0Zscore2, bs = "cr") + 
    s(durationZscore2, bs = "cr") + ti(measurement.no, f0Zscore2, 
    durationZscore2, bs = "cr") + ti(measurement.no, durationZscore2, 
    bs = "cr") + ti(measurement.no, f0Zscore2, bs = "cr") + ti(f0Zscore2, 
    durationZscore2, bs = "cr") + s(measurement.no, word, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1) + s(durationZscore2, 
    word, bs = "fs", xt = list(bs = "tp"), k = 3, m = 1) + s(f0Zscore2, 
    word, bs = "fs", xt = list(bs = "tp"), k = 3, m = 1) + s(measurement.no, 
    wordLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 3, 
    m = 1) + s(durationZscore2, wordLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 3, m = 1) + s(f0Zscore2, wordLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 3, m = 1) + s(measurement.no, 
    wordPos, bs = "fs", xt = list(bs = "tp"), k = 3, m = 1) + 
    s(durationZscore2, wordPos, bs = "fs", xt = list(bs = "tp"), 
        k = 3, m = 1) + s(f0Zscore2, wordPos, bs = "fs", xt = list(bs = "tp"), 
    k = 3, m = 1) + s(measurement.no, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + 
    s(f0Zscore2, speakerPos, bs = "fs", xt = list(bs = "tp"), 
        k = 10, m = 1) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 5, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 5, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 5, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.05456    0.03976  -1.372     0.17

Approximate significance of smooth terms:
                                               edf Ref.df      F  p-value    
s(measurement.no)                            7.999  8.165 70.719  < 2e-16 ***
s(f0Zscore2)                                 1.000  1.000  4.159   0.0414 *  
s(durationZscore2)                           2.014  2.094  0.911   0.3851    
ti(measurement.no,f0Zscore2,durationZscore2) 7.469 10.436  0.978   0.3984    
ti(measurement.no,durationZscore2)           8.978 10.828 18.751  < 2e-16 ***
ti(measurement.no,f0Zscore2)                 7.295  8.907  5.030 8.86e-07 ***
ti(f0Zscore2,durationZscore2)                3.516  4.304  1.386   0.2594    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.765   Deviance explained = 79.9%
fREML = 5915.7  Scale est. = 0.13714   n = 11029

gam.check(gamm.model2test)

Method: fREML   Optimizer: perf chol
$grad
 [1]  7.327472e-14 -5.359783e-05 -1.103562e-13 -1.332268e-14 -6.039613e-14 -2.509104e-14
 [7] -3.752554e-14  3.641532e-14  5.750955e-14 -1.971756e-13 -4.329870e-14 -1.409983e-14
[13]  2.842171e-14  1.509903e-13 -7.198618e-05  1.669775e-13 -9.287826e-05 -3.677059e-13
[19]  3.410605e-13 -1.154632e-13 -7.460699e-14  5.506706e-14  1.278977e-13  1.243450e-13
[25]  1.421085e-14 -1.119105e-13 -9.237056e-14 -6.927792e-14  3.126388e-13 -1.776357e-15
[31]  3.623768e-13 -1.845746e-15  7.105427e-15  1.666722e-14 -7.105427e-14  6.078471e-15
[37]  8.881784e-14 -5.018208e-14 -4.121148e-13 -1.021405e-14 -1.776357e-13  3.019807e-14
[43] -7.815970e-14  8.437695e-14  6.608047e-13  1.776357e-14  2.131628e-13 -1.438849e-13
[49] -1.818989e-12

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]
   3.515353e+00  1.734713e-07  6.250010e-05  2.099074e-03  1.067503e-03  2.136168e-03
   1.734713e-07  5.359434e-05 -6.617653e-08 -4.848540e-07 -2.724150e-07 -6.260771e-08
   6.250010e-05 -6.617653e-08  2.750637e-01 -1.144683e-03 -3.919240e-03  3.065156e-03
   2.099074e-03 -4.848540e-07 -1.144683e-03  3.764993e-01  5.267160e-02  5.381808e-02
   1.067503e-03 -2.724150e-07 -3.919240e-03  5.267160e-02  4.505835e-01 -1.278721e-01
   2.136168e-03 -6.260771e-08  3.065156e-03  5.381808e-02 -1.278721e-01  2.570803e-01
  -6.000821e-04 -6.604466e-08 -2.873264e-03 -1.537929e-02 -7.175089e-03  3.986232e-03
  -2.479263e-04  2.915065e-08 -2.390580e-04  5.205418e-04  5.834331e-04 -3.872980e-02
  -2.292334e-02 -1.067458e-06 -3.449031e-04  1.367769e-03 -2.245853e-03  2.736511e-03
  -2.625597e-03 -7.429757e-06 -3.257084e-04  3.639254e-02 -1.543300e-03 -1.852304e-02
  -3.172543e-03  4.152279e-07  8.924450e-03 -3.319540e-03 -8.346621e-03 -4.059068e-02
  -1.714842e-04  9.046096e-08 -2.473567e-02 -8.003946e-03 -1.491780e-02 -1.936009e-02
  -1.092248e-02 -6.981689e-07  5.732594e-03  8.519329e-03  5.162828e-02  5.466842e-02
  -2.682439e-04 -4.400221e-07 -7.630525e-03 -1.730844e-03  5.598155e-03  5.123250e-03
   4.000549e-07 -7.189309e-11  1.537678e-06 -5.278494e-07 -2.392877e-06 -1.403395e-06
  -2.682439e-04 -4.400221e-07 -7.630525e-03 -1.730844e-03  5.598155e-03  5.123250e-03
   1.069782e-07 -8.287089e-11 -5.593675e-07  2.551404e-07  1.883580e-07  5.319414e-07
  -2.682439e-04 -4.400221e-07 -7.630525e-03 -1.730844e-03  5.598155e-03  5.123250e-03
  -8.813946e-03 -2.042993e-06  1.011647e-03  5.638530e-02  4.375822e-02  4.065695e-02
  -3.718286e-04 -3.440895e-07 -7.182883e-03 -4.198553e-03  1.438183e-03  2.572849e-03
           [,7]          [,8]          [,9]         [,10]         [,11]         [,12]
  -6.000821e-04 -2.479263e-04 -2.292334e-02 -2.625597e-03 -3.172543e-03 -1.714842e-04
  -6.604466e-08  2.915065e-08 -1.067458e-06 -7.429757e-06  4.152279e-07  9.046096e-08
  -2.873264e-03 -2.390580e-04 -3.449031e-04 -3.257084e-04  8.924450e-03 -2.473567e-02
  -1.537929e-02  5.205418e-04  1.367769e-03  3.639254e-02 -3.319540e-03 -8.003946e-03
  -7.175089e-03  5.834331e-04 -2.245853e-03 -1.543300e-03 -8.346621e-03 -1.491780e-02
   3.986232e-03 -3.872980e-02  2.736511e-03 -1.852304e-02 -4.059068e-02 -1.936009e-02
   1.297922e+00 -4.721886e-02  3.555107e-03 -5.038050e-03 -9.403568e-03 -1.457051e-02
  -4.721886e-02  1.328429e+00 -1.277377e-02  1.859032e-03  2.421557e-02 -9.513306e-02
   3.555107e-03 -1.277377e-02  6.836243e-01  3.208912e-01 -1.700558e-03 -4.761105e-03
  -5.038050e-03  1.859032e-03  3.208912e-01  5.319962e-01  2.554781e-03 -1.201076e-02
  -9.403568e-03  2.421557e-02 -1.700558e-03  2.554781e-03  1.884662e-01  2.319193e-02
  -1.457051e-02 -9.513306e-02 -4.761105e-03 -1.201076e-02  2.319193e-02  2.977111e-01
   1.015023e-02  5.633674e-03 -1.196681e-02 -6.123377e-02  1.132118e-02  2.924093e-02
   1.204084e-04 -5.056881e-03  8.976213e-04  2.315556e-03 -2.873057e-04 -1.142670e-03
   8.609552e-07  1.057864e-06  5.367193e-07 -6.794350e-07  4.070293e-07 -5.908487e-06
   1.204084e-04 -5.056881e-03  8.976213e-04  2.315556e-03 -2.873057e-04 -1.142670e-03
   9.299422e-07 -1.121830e-06 -8.881177e-07 -2.514528e-06  3.351192e-06  1.548844e-06
   1.204084e-04 -5.056881e-03  8.976213e-04  2.315556e-03 -2.873057e-04 -1.142670e-03
   1.111674e-02 -6.637716e-03 -6.594935e-02  5.688982e-02  4.104975e-02  1.022953e-01
   1.568683e-04 -4.891373e-03 -6.254570e-04  6.202224e-03  1.113369e-03 -4.871156e-03
          [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -1.092248e-02 -2.682439e-04  4.000549e-07 -2.682439e-04  1.069782e-07 -2.682439e-04
  -6.981689e-07 -4.400221e-07 -7.189309e-11 -4.400221e-07 -8.287089e-11 -4.400221e-07
   5.732594e-03 -7.630525e-03  1.537678e-06 -7.630525e-03 -5.593675e-07 -7.630525e-03
   8.519329e-03 -1.730844e-03 -5.278494e-07 -1.730844e-03  2.551404e-07 -1.730844e-03
   5.162828e-02  5.598155e-03 -2.392877e-06  5.598155e-03  1.883580e-07  5.598155e-03
   5.466842e-02  5.123250e-03 -1.403395e-06  5.123250e-03  5.319414e-07  5.123250e-03
   1.015023e-02  1.204084e-04  8.609552e-07  1.204084e-04  9.299422e-07  1.204084e-04
   5.633674e-03 -5.056881e-03  1.057864e-06 -5.056881e-03 -1.121830e-06 -5.056881e-03
  -1.196681e-02  8.976213e-04  5.367193e-07  8.976213e-04 -8.881177e-07  8.976213e-04
  -6.123377e-02  2.315556e-03 -6.794350e-07  2.315556e-03 -2.514528e-06  2.315556e-03
   1.132118e-02 -2.873057e-04  4.070293e-07 -2.873057e-04  3.351192e-06 -2.873057e-04
   2.924093e-02 -1.142670e-03 -5.908487e-06 -1.142670e-03  1.548844e-06 -1.142670e-03
   9.039470e+00  2.754007e-03  2.615377e-06  2.754007e-03  1.092647e-05  2.754007e-03
   2.754007e-03  5.743026e-01  2.774251e-05  5.743026e-01  1.753174e-05  5.743026e-01
   2.615377e-06  2.774251e-05  7.203586e-05  2.774251e-05  5.069861e-10  2.774251e-05
   2.754007e-03  5.743026e-01  2.774251e-05  5.743026e-01  1.753174e-05  5.743026e-01
   1.092647e-05  1.753174e-05  5.069861e-10  1.753174e-05  9.289832e-05  1.753174e-05
   2.754007e-03  5.743026e-01  2.774251e-05  5.743026e-01  1.753174e-05  5.743026e-01
   1.522494e+01  5.765057e-02  2.158597e-05  5.765057e-02  5.313838e-05  5.765057e-02
   8.274888e-03  3.086136e-01  4.157147e-05  3.086136e-01  1.075506e-05  3.086136e-01
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]
  -8.813946e-03 -3.718286e-04  3.734551e-03 -3.718286e-04 -1.481476e-03 -3.718286e-04
  -2.042993e-06 -3.440895e-07 -6.999927e-07 -3.440895e-07 -3.244854e-06 -3.440895e-07
   1.011647e-03 -7.182883e-03 -2.998040e-02 -7.182883e-03 -1.800463e-02 -7.182883e-03
   5.638530e-02 -4.198553e-03 -8.990512e-03 -4.198553e-03  3.393670e-03 -4.198553e-03
   4.375822e-02  1.438183e-03 -8.285557e-03  1.438183e-03  2.517066e-02  1.438183e-03
   4.065695e-02  2.572849e-03  1.439072e-03  2.572849e-03  2.119409e-02  2.572849e-03
   1.111674e-02  1.568683e-04  1.232464e-02  1.568683e-04  3.335688e-02  1.568683e-04
  -6.637716e-03 -4.891373e-03 -3.182517e-02 -4.891373e-03  1.122041e-02 -4.891373e-03
  -6.594935e-02 -6.254570e-04  3.209375e-03 -6.254570e-04  6.935718e-04 -6.254570e-04
   5.688982e-02  6.202224e-03 -5.873285e-03  6.202224e-03 -5.413826e-03  6.202224e-03
   4.104975e-02  1.113369e-03  1.176798e-02  1.113369e-03  4.990998e-02  1.113369e-03
   1.022953e-01 -4.871156e-03 -6.961474e-02 -4.871156e-03  8.138101e-02 -4.871156e-03
   1.522494e+01  8.274888e-03  5.350253e-02  8.274888e-03  8.121419e-01  8.274888e-03
   5.765057e-02  3.086136e-01  1.770263e-01  3.086136e-01  2.910577e-01  3.086136e-01
   2.158597e-05  4.157147e-05  3.528778e-04  4.157148e-05  5.926741e-05  4.157148e-05
   5.765057e-02  3.086136e-01  1.770263e-01  3.086136e-01  2.910577e-01  3.086136e-01
   5.313838e-05  1.075506e-05 -6.904595e-06  1.075506e-05  2.210152e-04  1.075506e-05
   5.765057e-02  3.086136e-01  1.770263e-01  3.086136e-01  2.910577e-01  3.086136e-01
   5.823101e+01  4.128214e-02  2.965312e-01  4.128214e-02  2.250110e+00  4.128214e-02
   4.128214e-02  9.419166e-01  7.049264e-01  9.419166e-01  9.000911e-01  9.419166e-01
          [,25]         [,26]         [,27]         [,28]         [,29]         [,30]
   3.063532e-03 -1.227346e-03 -8.128050e-04 -1.227346e-03  4.761000e-04 -1.227346e-03
   5.512651e-07 -1.014388e-06 -4.376011e-07 -1.014388e-06 -4.287839e-06 -1.014388e-06
   1.017670e-02 -7.177007e-03  2.193903e-02 -7.177007e-03  2.367473e-04 -7.177007e-03
   2.059826e-02 -2.944719e-03 -6.173710e-04 -2.944719e-03  1.282408e-02 -2.944719e-03
   5.001594e-02  1.280059e-03 -2.773404e-02  1.280059e-03  3.029798e-02  1.280059e-03
   3.539306e-02  3.018760e-03 -1.370260e-02  3.018760e-03  2.336816e-02  3.018760e-03
   3.762737e-02  3.237817e-03  6.100248e-03  3.237817e-03  2.154058e-02  3.237817e-03
  -6.012762e-02 -6.610273e-03  1.841749e-02 -6.610273e-03  8.660457e-03 -6.610273e-03
   6.668586e-03 -5.908623e-04 -3.844081e-06 -5.908623e-04 -1.170313e-02 -5.908623e-04
   1.025079e-02  4.314305e-03 -9.438793e-03  4.314305e-03 -4.890537e-02  4.314305e-03
   5.997339e-03  8.798648e-03  3.918751e-03  8.798648e-03  9.076909e-02  8.798648e-03
   2.470996e-02  3.706717e-03 -4.040108e-02  3.706717e-03  3.568227e-02  3.706717e-03
   3.247231e+00 -8.331614e-03  8.495592e-02 -8.331614e-03  4.095826e-01 -8.331614e-03
  -3.865610e-03  4.380561e-01  3.288633e-01  4.380561e-01  3.091767e-01  4.380561e-01
   2.376135e-06  2.417486e-05  3.438198e-04  2.417486e-05  2.468378e-05  2.417486e-05
  -3.865610e-03  4.380561e-01  3.288633e-01  4.380561e-01  3.091767e-01  4.380561e-01
   2.419183e-05  2.132564e-05  3.593498e-05  2.132564e-05  2.564578e-04  2.132564e-05
  -3.865610e-03  4.380561e-01  3.288633e-01  4.380561e-01  3.091767e-01  4.380561e-01
   8.009087e+00  5.012105e-02  3.008885e-01  5.012105e-02  1.802129e+00  5.012105e-02
   3.592641e-02  2.451223e-01  5.351499e-01  2.451223e-01  1.306684e-01  2.451223e-01
          [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -1.957419e-02  8.651916e-05 -4.634872e-04  8.651916e-05  1.083711e-03  8.651916e-05
  -6.090473e-07  7.013495e-09  7.869128e-08  7.013495e-09 -4.905629e-06  7.013495e-09
   1.078264e-03 -3.361113e-04  1.892776e-02 -3.361113e-04 -1.205936e-02 -3.361113e-04
   7.127771e-02 -3.077114e-04 -1.929128e-03 -3.077114e-04  9.572064e-03 -3.077114e-04
   2.785233e-02 -7.086155e-05 -4.265287e-03 -7.086155e-05  9.048170e-03 -7.086155e-05
   2.391496e-03 -1.774993e-04 -7.341357e-03 -1.774993e-04  1.547170e-02 -1.774993e-04
  -2.139034e-02  8.559990e-05 -8.864555e-04  8.559990e-05  3.989537e-03  8.559990e-05
   3.040254e-02  1.562156e-04 -7.282298e-03  1.562156e-04  3.009222e-02  1.562156e-04
  -4.584941e-02  2.482773e-04  1.914068e-03  2.482773e-04 -6.792545e-03  2.482773e-04
   4.921107e-02 -1.700974e-04  7.715210e-04 -1.700974e-04 -1.584932e-01 -1.700974e-04
   1.280202e-04 -9.130059e-05  9.164045e-03 -9.130059e-05  2.203064e-02 -9.130060e-05
   9.115963e-03  1.110850e-03 -1.232866e-02  1.110850e-03  6.645245e-03  1.110850e-03
   1.345234e-01 -1.716492e-04  1.563964e-02 -1.716492e-04  1.261459e-02 -1.716492e-04
  -4.506500e-03 -2.992545e-05  1.108883e-02 -2.992545e-05 -1.350212e-02 -2.992545e-05
   8.228938e-06 -4.163405e-07  1.584191e-05 -4.163405e-07 -1.346311e-05 -4.163405e-07
  -4.506500e-03 -2.992545e-05  1.108883e-02 -2.992545e-05 -1.350212e-02 -2.992545e-05
   4.077058e-06 -7.221842e-08  3.464073e-06 -7.221842e-08 -4.635750e-06 -7.221842e-08
  -4.506500e-03 -2.992545e-05  1.108883e-02 -2.992545e-05 -1.350212e-02 -2.992545e-05
   2.417953e-01 -1.137978e-03 -1.566450e-02 -1.137978e-03 -1.946942e-02 -1.137978e-03
  -1.309974e-03 -1.997786e-03  7.995700e-03 -1.997786e-03 -5.161172e-02 -1.997786e-03
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]
   6.508631e-03  4.542543e-04 -5.200397e-04  4.542543e-04  2.331807e-03  4.542543e-04
  -1.481997e-06  1.017547e-07 -4.653770e-07  1.017547e-07  2.056746e-07  1.017547e-07
   9.050014e-03  6.535860e-03  3.845691e-03  6.535860e-03  1.043818e-02  6.535860e-03
  -5.100845e-02 -1.665356e-03  1.088982e-02 -1.665356e-03  1.752515e-02 -1.665356e-03
   4.876315e-04  3.073109e-04  1.449638e-02  3.073109e-04 -5.928034e-03  3.073109e-04
  -1.381394e-02  2.181238e-03  7.266178e-04  2.181238e-03  5.931124e-03  2.181238e-03
  -1.014863e-02  1.127687e-04  3.653871e-03  1.127687e-04  7.134118e-03  1.127687e-04
  -1.436441e-02 -2.515113e-03 -3.619174e-02 -2.515113e-03  9.361054e-02 -2.515113e-03
   2.435898e-03  6.234950e-04  6.335894e-03  6.234950e-04  2.423515e-02  6.234950e-04
  -5.356073e-02  2.237072e-03  2.400343e-02  2.237072e-03  2.064235e-02  2.237072e-03
  -2.704586e-02 -9.774293e-04 -3.355778e-02 -9.774293e-04  1.467168e-02 -9.774293e-04
  -2.949337e-02 -3.283299e-03 -7.225536e-02 -3.283299e-03  9.286753e-03 -3.283299e-03
   1.175207e-01 -1.435919e-03 -9.137602e-02 -1.435919e-03  3.261554e-02 -1.435919e-03
   1.423516e-02 -8.489582e-03 -5.166855e-02 -8.489582e-03 -3.400680e-02 -8.489582e-03
   3.288978e-06 -3.087494e-06  8.675636e-05 -3.087494e-06  9.526133e-07 -3.087494e-06
   1.423516e-02 -8.489582e-03 -5.166855e-02 -8.489582e-03 -3.400680e-02 -8.489582e-03
   4.282721e-06 -1.347577e-06 -1.246200e-05 -1.347577e-06  1.956037e-06 -1.347577e-06
   1.423516e-02 -8.489582e-03 -5.166855e-02 -8.489582e-03 -3.400680e-02 -8.489582e-03
  -2.688123e-01 -7.374267e-03 -3.004114e-01 -7.374267e-03 -2.070009e-01 -7.374267e-03
   1.486904e-02 -7.306829e-03  3.146466e-02 -7.306829e-03 -4.575714e-02 -7.306829e-03
          [,43]         [,44]         [,45]         [,46]         [,47]         [,48]
   1.629493e-02  9.412544e-04 -5.310086e-03  9.412544e-04  2.058337e-03  9.412544e-04
  -4.765144e-07 -2.221551e-07 -1.491835e-08 -2.221551e-07 -1.559052e-06 -2.221551e-07
  -6.180670e-04 -2.805909e-03  4.966234e-02 -2.805909e-03  1.025767e-02 -2.805909e-03
  -4.158215e-02 -4.130472e-03  9.927497e-03 -4.130472e-03 -2.465117e-02 -4.130472e-03
  -4.346989e-03  1.229781e-03 -3.850640e-03  1.229781e-03 -2.551288e-03  1.229781e-03
   1.035366e-02  1.839256e-03  1.727775e-02  1.839256e-03 -1.552691e-02  1.839256e-03
   2.945919e-02  8.956551e-05 -8.428906e-03  8.956552e-05 -3.195256e-03  8.956551e-05
   7.288848e-02 -5.108158e-03 -3.615952e-02 -5.108158e-03  1.686253e-02 -5.108158e-03
  -6.918040e-02  1.664197e-03 -8.341378e-03  1.664197e-03  9.002733e-03  1.664197e-03
   1.926792e-01 -4.996207e-03  6.468193e-06 -4.996207e-03 -2.675100e-03 -4.996207e-03
  -2.623495e-02  5.656942e-03  2.688190e-02  5.656942e-03 -2.870739e-02  5.656942e-03
   4.304534e-02  5.365013e-03 -2.682179e-02  5.365013e-03 -1.802384e-02  5.365013e-03
   9.330805e-01  8.486932e-03  1.014504e-01  8.486932e-03  4.393746e-02  8.486932e-03
  -1.796420e-02  3.683902e-02  1.138147e-02  3.683902e-02  3.409380e-02  3.683902e-02
  -3.906305e-06 -1.105539e-06  5.548511e-06 -1.105539e-06 -1.428680e-05 -1.105539e-06
  -1.796420e-02  3.683902e-02  1.138147e-02  3.683902e-02  3.409380e-02  3.683902e-02
   2.946578e-06  3.387541e-06 -4.474957e-06  3.387541e-06  4.161140e-06  3.387541e-06
  -1.796420e-02  3.683902e-02  1.138147e-02  3.683902e-02  3.409380e-02  3.683902e-02
   4.292433e+00  6.159721e-02  3.127001e-01  6.159721e-02  1.781564e-01  6.159721e-02
  -4.141968e-02  1.202989e-01  9.683877e-02  1.202989e-01  3.063189e-02  1.202989e-01
          [,49]
  -3.499505e+00
  -2.542367e-05
  -5.070718e-01
  -1.176987e+00
  -1.153294e+00
  -9.042403e-01
  -1.688430e+00
  -2.300393e+00
  -1.272580e+00
  -1.875046e+00
  -5.905415e-01
  -6.672805e-01
  -4.170095e+01
  -6.711692e+00
  -2.315759e-03
  -6.711692e+00
  -1.613599e-03
  -6.711692e+00
  -1.346091e+02
  -1.068569e+01
 [ getOption("max.print") est atteint -- 29 lignes omises ]

Model rank =  9875 / 9875 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                 k'    edf k-index p-value    
s(measurement.no)                              9.00   8.00    0.98    0.14    
s(f0Zscore2)                                   9.00   1.00    1.00    0.51    
s(durationZscore2)                             9.00   2.01    0.94  <2e-16 ***
ti(measurement.no,f0Zscore2,durationZscore2)  64.00   7.47    1.00    0.40    
ti(measurement.no,durationZscore2)            16.00   8.98    0.99    0.17    
ti(measurement.no,f0Zscore2)                  16.00   7.30    1.00    0.61    
ti(f0Zscore2,durationZscore2)                 16.00   3.52    0.95  <2e-16 ***
s(measurement.no,word)                       627.00  96.83    0.98    0.09 .  
s(durationZscore2,word)                      627.00  13.43    0.94  <2e-16 ***
s(f0Zscore2,word)                            627.00  13.43    1.00    0.52    
s(measurement.no,wordLeftRightTone)          930.00 290.59    0.98    0.10    
s(durationZscore2,wordLeftRightTone)         930.00  83.88    0.94  <2e-16 ***
s(f0Zscore2,wordLeftRightTone)               930.00 126.58    1.00    0.52    
s(measurement.no,wordPos)                    993.00 112.59    0.98    0.07 .  
s(durationZscore2,wordPos)                   993.00 101.30    0.94  <2e-16 ***
s(f0Zscore2,wordPos)                         993.00 134.51    1.00    0.51    
s(measurement.no,speaker)                    100.00  65.87    0.98    0.14    
s(durationZscore2,speaker)                   100.00  13.17    0.94  <2e-16 ***
s(f0Zscore2,speaker)                         100.00  42.19    1.00    0.51    
s(measurement.no,speakerPos)                 300.00  42.17    0.98    0.07 .  
s(durationZscore2,speakerPos)                300.00 117.44    0.94  <2e-16 ***
s(f0Zscore2,speakerPos)                      300.00  59.01    1.00    0.52    
s(measurement.no,speakerLeftRightTone)       295.00  82.70    0.98    0.09 .  
s(durationZscore2,speakerLeftRightTone)      295.00  95.50    0.94  <2e-16 ***
s(f0Zscore2,speakerLeftRightTone)            295.00  48.67    1.00    0.48    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Because of the high number of levels in the word variable — for instance, in the case of the diphthong /ai/ produced by male speakers, there were 205 distinct word types, each appearing on average only five times. The related variables wordPos and wordLeftRightTone contained 311 and 293 levels, respectively. Introducing separate nonlinear effects for each of these rarely occurring levels and increasing k values, would increase model fitting, but at the cost of substantially greater computation time. Moreover, simply doubling the k values failed to resolve the poor fit, as reflected in the gam.check diagnostics.

Another structure is tested:

5.2 Model 2A: /ai/ male, F1 as output

system.time(gamm.model2a.noAR <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2a.noAR, paste("Gamm_model2a_noAR.rds"))

gamm.model2a.noAR <- 
  readRDS("Gamm_model2a_noAR.rds")

r.gamm.model2a <- start_value_rho(gamm.model2a.noAR)

system.time(gamm.model2a <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
                      
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2a, AR.start = data.ai.mas$start)
)

#saveRDS(gamm.model2a, paste("Gamm_model2a.rds"))

gamm.model2a <- 
  readRDS("Gamm_model2a.rds")

summary(gamm.model2a, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.04726    0.04994  -0.946    0.344

Approximate significance of smooth terms:
                                                edf Ref.df      F  p-value    
s(measurement.no)                             7.931  8.103 78.102  < 2e-16 ***
s(f0Zscore2)                                  1.000  1.000  2.757   0.0969 .  
s(durationZscore2)                            1.883  1.926  0.843   0.4655    
ti(measurement.no,f0Zscore2,durationZscore2) 39.275 58.289  1.137   0.2213    
ti(measurement.no,durationZscore2)            8.530 10.584 17.884  < 2e-16 ***
ti(measurement.no,f0Zscore2)                 10.303 13.169  3.641 7.66e-06 ***
ti(f0Zscore2,durationZscore2)                 6.803  9.395  0.817   0.6130    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.773   Deviance explained = 80.5%
fREML = 6007.6  Scale est. = 0.13386   n = 11029

gam.check(gamm.model2a)

Method: fREML   Optimizer: perf chol
$grad
 [1]  5.462297e-14 -9.626245e-05 -5.856648e-12  2.785328e-12  5.072387e-12  2.330580e-12 -7.629453e-13  1.047162e-12  3.013145e-13
[10]  9.966250e-12  5.706546e-14  6.866285e-12 -4.843059e-11 -1.431744e-12 -6.084733e-11  6.449596e-11  6.220446e-11 -9.343033e-10
[19]  1.883365e-10  1.088480e-10 -2.186091e-09  2.705391e-11 -2.397638e-11 -7.213591e-05  6.110668e-12 -7.087024e-05 -7.877698e-11
[28] -7.087024e-05  1.065814e-14 -7.087024e-05 -2.700062e-13  1.106071e-11 -1.008971e-10  1.121658e-11 -7.982948e-12  1.112643e-11
[37] -2.032152e-12  6.528111e-12  4.175149e-11  6.549428e-12  9.016787e-12  6.595169e-12  4.829417e-09

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.466488e+00  3.254244e-07 -5.257459e-05  2.369806e-03 -3.785818e-04  9.031393e-04 -8.953825e-04  2.513386e-03 -2.314106e-02
   3.254244e-07  9.624767e-05  2.522700e-08 -1.672270e-06  3.218722e-07 -3.379400e-07 -3.567809e-07  1.001399e-07 -1.830944e-06
  -5.257459e-05  2.522700e-08  2.428702e-01  9.599315e-03 -3.791693e-03 -4.473523e-04 -2.250122e-03  4.141712e-03 -3.018283e-04
   2.369806e-03 -1.672270e-06  9.599315e-03  1.491842e+00  3.739066e-01  3.498922e-01 -1.445288e-02  8.733083e-03 -1.701172e-02
  -3.785818e-04  3.218722e-07 -3.791693e-03  3.739066e-01  1.039836e+00  3.620979e-01 -4.001896e-02  1.331440e-02  1.315308e-02
   9.031393e-04 -3.379400e-07 -4.473523e-04  3.498922e-01  3.620979e-01  1.665344e+00 -2.164189e-02  5.687499e-02 -1.384919e-02
  -8.953825e-04 -3.567809e-07 -2.250122e-03 -1.445288e-02 -4.001896e-02 -2.164189e-02  1.312628e+00 -4.818861e-02  6.458888e-03
   2.513386e-03  1.001399e-07  4.141712e-03  8.733083e-03  1.331440e-02  5.687499e-02 -4.818861e-02  7.147504e-01 -1.097371e-02
  -2.314106e-02 -1.830944e-06 -3.018283e-04 -1.701172e-02  1.315308e-02 -1.384919e-02  6.458888e-03 -1.097371e-02  8.211776e-01
   8.670701e-04 -1.135610e-05  2.368823e-04  1.868331e-02 -2.159962e-02 -2.776271e-02 -3.243167e-03  8.905405e-03  2.849521e-01
  -2.481418e-03  7.630919e-07  1.228556e-02  9.108747e-03 -2.231992e-02 -4.294433e-02 -7.035073e-03  1.259923e-02 -1.893175e-03
   2.556338e-04  8.581353e-09 -2.427160e-02 -5.468547e-03 -7.785009e-02 -4.163480e-02 -1.012262e-02 -4.430971e-02 -2.490084e-03
  -3.447358e-03  7.005046e-07 -2.786481e-02  6.451613e-02  8.385667e-02  1.940479e-01  7.534175e-03  6.382948e-03 -1.940439e-02
  -1.803509e-03 -1.808460e-06 -8.122175e-03  1.695652e-02  1.476343e-02  5.458842e-02  1.419975e-02 -1.016646e-02  1.234741e-03
  -3.135282e-03  1.014155e-06 -4.549273e-03 -7.143607e-03 -1.932942e-02 -1.324133e-02  7.630295e-03  1.044614e-02  1.414855e-03
   3.161807e-03  1.948018e-06 -2.931742e-02 -1.547613e-02 -3.248714e-02  3.228465e-02  9.935850e-03  1.245089e-02  6.326217e-03
  -7.303350e-05 -3.552654e-06  4.711371e-03  4.852834e-02  9.054060e-04  1.546967e-02  2.733553e-03 -2.295500e-02 -4.355489e-03
  -2.510954e-03 -8.461292e-07 -9.895568e-03 -8.636499e-03 -3.037666e-02  1.923186e-02  2.530829e-03  5.492548e-03 -7.789324e-03
  -1.146528e-02  2.288220e-09  5.695517e-03 -1.089927e-02 -2.148931e-02  1.248538e-01  9.588392e-03  3.768874e-02  2.733393e-03
  -9.108057e-04 -8.447339e-06 -1.360835e-02  5.705969e-02  4.784689e-02  2.705386e-02  1.881434e-02 -1.652159e-02  3.934800e-03
   1.285633e-03 -4.643584e-07 -3.077620e-02 -5.300816e-02 -3.104807e-02 -6.271712e-03  8.709306e-03 -1.461632e-02 -2.470352e-03
  -8.029424e-04  1.473774e-07  1.398889e-03  8.398311e-04  8.201404e-04  7.450023e-02  3.883319e-03 -8.474142e-03 -1.166961e-03
   5.296443e-04 -9.827522e-07 -4.296249e-04  3.680081e-03  1.541388e-04 -2.672043e-04  1.606137e-03 -4.885811e-03 -1.162341e-03
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   8.670701e-04 -2.481418e-03  2.556338e-04 -3.447358e-03 -1.803509e-03 -3.135282e-03  3.161807e-03 -7.303350e-05 -2.510954e-03
  -1.135610e-05  7.630919e-07  8.581353e-09  7.005046e-07 -1.808460e-06  1.014155e-06  1.948018e-06 -3.552654e-06 -8.461292e-07
   2.368823e-04  1.228556e-02 -2.427160e-02 -2.786481e-02 -8.122175e-03 -4.549273e-03 -2.931742e-02  4.711371e-03 -9.895568e-03
   1.868331e-02  9.108747e-03 -5.468547e-03  6.451613e-02  1.695652e-02 -7.143607e-03 -1.547613e-02  4.852834e-02 -8.636499e-03
  -2.159962e-02 -2.231992e-02 -7.785009e-02  8.385667e-02  1.476343e-02 -1.932942e-02 -3.248714e-02  9.054060e-04 -3.037666e-02
  -2.776271e-02 -4.294433e-02 -4.163480e-02  1.940479e-01  5.458842e-02 -1.324133e-02  3.228465e-02  1.546967e-02  1.923186e-02
  -3.243167e-03 -7.035073e-03 -1.012262e-02  7.534175e-03  1.419975e-02  7.630295e-03  9.935850e-03  2.733553e-03  2.530829e-03
   8.905405e-03  1.259923e-02 -4.430971e-02  6.382948e-03 -1.016646e-02  1.044614e-02  1.245089e-02 -2.295500e-02  5.492548e-03
   2.849521e-01 -1.893175e-03 -2.490084e-03 -1.940439e-02  1.234741e-03  1.414855e-03  6.326217e-03 -4.355489e-03 -7.789324e-03
   7.285349e-01  1.011165e-04 -1.010159e-02 -1.070318e-02 -4.019067e-02  5.062560e-02  4.234568e-03  1.364799e-02 -2.422853e-02
   1.011165e-04  2.595572e-01  3.861788e-02  1.179314e-02  2.839956e-02  7.204392e-03  3.833444e-02  4.256016e-02 -2.560959e-03
  -1.010159e-02  3.861788e-02  5.545796e-01  1.186113e-02  1.839866e-02  1.115068e-02  3.516205e-02 -2.483921e-02 -3.471128e-02
  -1.070318e-02  1.179314e-02  1.186113e-02  1.815138e+01  4.309702e+00  3.947410e+00 -7.340032e-01 -5.992768e-02  7.390915e-01
  -4.019067e-02  2.839956e-02  1.839866e-02  4.309702e+00  9.218066e+00  8.806554e-01  5.940487e-01  4.399999e-01  2.947558e-01
   5.062560e-02  7.204392e-03  1.115068e-02  3.947410e+00  8.806554e-01  7.929611e+00  8.904016e-02  1.155338e-01  3.394582e-01
   4.234568e-03  3.833444e-02  3.516205e-02 -7.340032e-01  5.940487e-01  8.904016e-02  1.435087e+01  4.821219e-01 -3.392921e-01
   1.364799e-02  4.256016e-02 -2.483921e-02 -5.992768e-02  4.399999e-01  1.155338e-01  4.821219e-01  7.455341e+00  1.166365e-01
  -2.422853e-02 -2.560959e-03 -3.471128e-02  7.390915e-01  2.947558e-01  3.394582e-01 -3.392921e-01  1.166365e-01  3.660416e+00
   9.722005e-02  4.191458e-02  3.337161e-02 -3.128248e+00 -7.502600e-01 -2.509511e+00  9.658708e+00  1.898687e-01 -4.095034e-01
  -3.582614e-02  8.281039e-02  7.335226e-02  7.455916e-02 -5.440680e-02 -3.370936e-01  8.115281e-02  3.156737e+00 -2.460427e-01
  -3.958237e-02 -1.626170e-02 -6.262443e-02  5.366123e-01  3.051813e-01  6.735285e-01 -3.297105e-01 -2.801922e-02  1.747584e+00
  -1.100917e-02  1.028274e-02  1.175056e-02 -1.162598e+00 -2.267117e-01 -2.644962e-01  2.589457e+00 -2.963317e-02 -8.849255e-02
  -6.160560e-03  5.084718e-03  2.038801e-03  1.068396e-02  5.458985e-03  1.597971e-02  1.581451e-02  4.699794e-01  2.319510e-02
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
  -1.146528e-02 -9.108057e-04  1.285633e-03 -8.029424e-04  5.296443e-04  1.902710e-07 -6.273515e-03  3.238994e-07  2.256300e-04
   2.288220e-09 -8.447339e-06 -4.643584e-07  1.473774e-07 -9.827522e-07 -5.189064e-11 -1.628522e-06  1.978558e-10  9.811981e-08
   5.695517e-03 -1.360835e-02 -3.077620e-02  1.398889e-03 -4.296249e-04 -1.342548e-06  2.765686e-04 -1.068288e-06  5.185318e-03
  -1.089927e-02  5.705969e-02 -5.300816e-02  8.398311e-04  3.680081e-03 -1.676807e-06  5.710830e-02 -1.078327e-06 -1.962422e-03
  -2.148931e-02  4.784689e-02 -3.104807e-02  8.201404e-04  1.541388e-04 -1.239502e-06  4.982816e-02  8.550915e-08 -1.802298e-03
   1.248538e-01  2.705386e-02 -6.271712e-03  7.450023e-02 -2.672043e-04  9.230099e-07 -1.492911e-02  7.205055e-07 -2.824205e-03
   9.588392e-03  1.881434e-02  8.709306e-03  3.883319e-03  1.606137e-03  4.034183e-07 -5.019764e-03  4.341406e-07  3.961336e-04
   3.768874e-02 -1.652159e-02 -1.461632e-02 -8.474142e-03 -4.885811e-03 -2.500351e-07  3.256103e-02 -3.121217e-07 -1.117035e-03
   2.733393e-03  3.934800e-03 -2.470352e-03 -1.166961e-03 -1.162341e-03 -2.486333e-07 -4.118656e-02  1.391181e-06  1.336206e-03
   9.722005e-02 -3.582614e-02 -3.958237e-02 -1.100917e-02 -6.160560e-03  1.801658e-07 -8.338279e-03 -4.885601e-07  3.107099e-03
   4.191458e-02  8.281039e-02 -1.626170e-02  1.028274e-02  5.084718e-03  8.475368e-07  5.612739e-03  1.868643e-06  8.702710e-03
   3.337161e-02  7.335226e-02 -6.262443e-02  1.175056e-02  2.038801e-03 -1.554367e-06  9.990891e-03  7.728639e-06  2.576811e-03
  -3.128248e+00  7.455916e-02  5.366123e-01 -1.162598e+00  1.068396e-02  3.280058e-05 -7.674887e-02  4.688782e-07  1.097912e-01
  -7.502600e-01 -5.440680e-02  3.051813e-01 -2.267117e-01  5.458985e-03  1.524673e-05 -3.286473e-02  3.958571e-06  2.221275e-02
  -2.509511e+00 -3.370936e-01  6.735285e-01 -2.644962e-01  1.597971e-02  1.837905e-05 -8.305300e-03 -1.474733e-05  4.282762e-02
   9.658708e+00  8.115281e-02 -3.297105e-01  2.589457e+00  1.581451e-02 -1.630738e-05  8.252216e-02  3.392793e-06 -1.420215e-03
   1.898687e-01  3.156737e+00 -2.801922e-02 -2.963317e-02  4.699794e-01  7.455051e-06  4.648056e-02 -8.819364e-06 -9.772582e-03
  -4.095034e-01 -2.460427e-01  1.747584e+00 -8.849255e-02  2.319510e-02  1.211135e-04 -3.012453e-02  3.493841e-06 -2.479951e-02
   4.479019e+01 -2.567000e-01 -3.915212e-01  7.181289e+00  3.549589e-02 -2.257985e-05  3.013539e-01 -1.866720e-06 -6.566352e-02
  -2.567000e-01  7.572756e+00 -3.249392e-01  1.634742e-01  5.473824e-01 -1.565755e-05  1.571203e-01 -4.565630e-06 -1.715742e-02
  -3.915212e-01 -3.249392e-01  3.631740e+00 -8.486202e-02 -2.279671e-02  1.354024e-04  4.393043e-02 -1.283146e-05 -1.239498e-02
   7.181289e+00  1.634742e-01 -8.486202e-02  2.751956e+00 -1.176012e-02 -5.099138e-06  8.718836e-02  1.070211e-06 -6.691574e-03
   3.549589e-02  5.473824e-01 -2.279671e-02 -1.176012e-02  1.327961e-01 -1.184830e-06  2.578065e-02 -5.003558e-08  1.953858e-03
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
   3.238994e-07  2.848656e-04  3.238994e-07  2.354045e-03  6.602446e-04  2.066015e-03  6.602446e-04  2.954308e-03  6.602446e-04
   1.978558e-10 -5.075100e-06  1.978558e-10 -2.357042e-06  1.742967e-07 -4.506694e-07  1.742967e-07  1.982060e-06  1.742967e-07
  -1.068288e-06 -1.060609e-02 -1.068288e-06  7.408756e-03  3.384804e-03  5.053837e-02  3.384804e-03  1.011859e-03  3.384804e-03
  -1.078327e-06  5.281122e-02 -1.078327e-06 -6.252265e-02 -6.166565e-03 -7.175672e-02 -6.166565e-03  2.905527e-02 -6.166565e-03
   8.550915e-08 -4.717364e-02  8.550915e-08  5.092826e-02 -1.270054e-02  6.507221e-02 -1.270054e-02 -6.738047e-02 -1.270054e-02
   7.205055e-07  2.785336e-02  7.205055e-07 -2.272129e-02  2.786613e-03 -5.384258e-04  2.786613e-03  3.737048e-02  2.786613e-03
   4.341406e-07 -6.816300e-04  4.341406e-07 -4.589540e-03  3.008454e-04  9.131853e-03  3.008454e-04 -5.655566e-03  3.008454e-04
  -3.121217e-07  3.067706e-02 -3.121217e-07 -8.937439e-03 -2.356265e-03 -1.194732e-02 -2.356265e-03  2.393652e-02 -2.356265e-03
   1.391181e-06 -8.951553e-04  1.391181e-06  2.654937e-02  1.297923e-03  9.785086e-03  1.297923e-03  3.370070e-02  1.297923e-03
  -4.885601e-07 -1.217557e-01 -4.885601e-07 -5.539353e-02  3.635620e-03  4.574992e-02  3.635620e-03  3.774287e-03  3.635620e-03
   1.868643e-06  2.808647e-02  1.868643e-06 -8.895873e-03  2.822619e-04  2.201661e-02  2.822619e-04 -4.660109e-03  2.822619e-04
   7.728639e-06 -1.411400e-02  7.728639e-06 -1.532168e-03 -2.168444e-03  2.031596e-02 -2.168444e-03 -1.685863e-02 -2.168444e-03
   4.688782e-07  2.020178e-02  4.688782e-07  3.589547e-01 -3.851288e-02  5.405672e-01 -3.851288e-02  1.720583e-01 -3.851288e-02
   3.958571e-06 -2.647148e-02  3.958571e-06  1.827308e-01 -4.478051e-03  3.656100e-01 -4.478051e-03 -2.230772e-01 -4.478051e-03
  -1.474733e-05 -4.780749e-02 -1.474733e-05  7.950423e-02 -1.680504e-02  5.065197e-01 -1.680504e-02  7.661485e-02 -1.680504e-02
   3.392793e-06 -3.648350e-02  3.392793e-06 -1.453652e-01 -7.329776e-03 -2.339512e-01 -7.329776e-03 -1.835476e-01 -7.329776e-03
  -8.819364e-06  4.519786e-02 -8.819364e-06  7.703553e-03 -4.630002e-02 -4.444761e-01 -4.630002e-02 -4.427686e-01 -4.630002e-02
   3.493841e-06 -3.832558e-02  3.493841e-06 -5.269946e-02 -3.825081e-02  1.309363e-01 -3.825081e-02  3.935735e-02 -3.825081e-02
  -1.866720e-06 -7.874127e-02 -1.866720e-06 -3.135266e-01  7.528121e-03 -8.271734e-01  7.528121e-03 -3.995199e-01  7.528121e-03
  -4.565630e-06 -2.562267e-01 -4.565630e-06  9.679624e-02 -3.790972e-02  8.421184e-02 -3.790972e-02  4.275194e-02 -3.790972e-02
  -1.283146e-05  1.860514e-02 -1.283146e-05 -4.356539e-02 -9.766193e-03  1.326428e-01 -9.766193e-03  1.783945e-01 -9.766193e-03
   1.070211e-06 -1.684891e-02  1.070211e-06 -9.637297e-03 -3.159678e-03 -2.009227e-01 -3.159679e-03 -1.318745e-02 -3.159679e-03
  -5.003558e-08  1.217378e-03 -5.003558e-08  1.799551e-02 -5.047812e-03 -1.458993e-02 -5.047812e-03  1.677832e-02 -5.047812e-03
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]         [,43]
  -3.213077e-04  3.019502e-04  3.543897e-04  3.019502e-04  1.260611e-03  3.019502e-04 -3.465615e+00
  -1.633042e-06  1.494054e-07 -2.300773e-06  1.494054e-07  2.556078e-06  1.494054e-07 -5.939244e-05
  -1.056913e-02 -3.745522e-03  7.688511e-02 -3.745522e-03  1.040335e-02 -3.745522e-03 -4.414147e-01
  -3.120067e-02 -3.601232e-03  1.183837e-02 -3.601232e-03 -4.018379e-02 -3.601232e-03 -6.506775e+00
   6.241151e-02 -6.620320e-04  2.134548e-02 -6.620320e-04 -9.395151e-03 -6.620320e-04 -5.662798e+00
   5.190466e-02 -2.014853e-03  6.995841e-02 -2.014853e-03 -1.517668e-02 -2.014853e-03 -6.967924e+00
   2.282296e-02  5.745013e-04  2.916525e-04  5.745013e-04 -1.075087e-02  5.745013e-04 -1.609863e+00
   1.643976e-02  1.132104e-04  1.316086e-02  1.132104e-04 -9.210980e-04  1.132104e-04 -2.155326e+00
  -7.294338e-02  1.354866e-03 -7.389988e-04  1.354866e-03  9.068532e-03  1.354866e-03 -1.549261e+00
   1.773028e-01 -1.506083e-03  1.538220e-02 -1.506083e-03 -2.658573e-02 -1.506083e-03 -3.102400e+00
   2.657577e-02  9.019459e-03  1.313715e-01  9.019459e-03 -1.934220e-02  9.019459e-03 -1.426948e+00
   5.511488e-02  1.267216e-02  1.173007e-02  1.267216e-02  3.694716e-02  1.267216e-02 -1.474379e+00
   1.556021e-01  1.877833e-02  1.306755e+00  1.877833e-02  8.748383e-02  1.877833e-02 -4.039977e+01
  -2.282520e-02  6.587446e-03  7.848954e-01  6.587446e-03 -1.297808e-02  6.587446e-03 -2.980972e+01
   2.759076e-01 -7.007738e-02  4.252907e-01 -7.007738e-02 -1.508098e-01 -7.007738e-02 -2.781311e+01
   7.613142e-01 -2.667143e-02 -2.427464e-01 -2.667143e-02 -9.270626e-02 -2.667143e-02 -4.495953e+01
   3.308783e-01 -2.499769e-03 -1.880997e-01 -2.499769e-03  1.798947e-01 -2.499769e-03 -2.692800e+01
   2.706501e-02 -7.299993e-02  4.475465e-01 -7.299993e-02 -1.829990e-02 -7.299993e-02 -1.401381e+01
   2.646707e+00  8.122295e-02 -9.287438e-01  8.122295e-02  4.696571e-02  8.122295e-02 -8.483872e+01
   9.535406e-01 -1.968214e-02 -3.232926e-01 -1.968214e-02  8.340616e-01 -1.968214e-02 -3.915445e+01
   2.194926e-04 -5.474025e-02  4.690387e-01 -5.474025e-02 -1.067795e-01 -5.474025e-02 -1.575472e+01
   5.212558e-01  9.763909e-03 -6.007923e-02  9.763909e-03  2.774211e-02  9.763909e-03 -1.671337e+01
   1.495984e-02 -7.913897e-03 -2.025652e-02 -7.913897e-03  1.910262e-02 -7.913897e-03 -2.896180e+00
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  8855 / 8855 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value
s(measurement.no)                            9.00e+00 7.93e+00    1.05    1.00
s(f0Zscore2)                                 9.00e+00 1.00e+00    1.00    0.56
s(durationZscore2)                           2.90e+01 1.88e+00    1.00    0.42
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 3.93e+01    1.01    0.86
ti(measurement.no,durationZscore2)           4.40e+01 8.53e+00    0.99    0.28
ti(measurement.no,f0Zscore2)                 2.80e+01 1.03e+01    0.99    0.18
ti(f0Zscore2,durationZscore2)                7.70e+01 6.80e+00    0.99    0.29
s(word)                                      2.09e+02 8.08e+01      NA      NA
s(wordPos)                                   3.31e+02 5.96e+01      NA      NA
s(wordLeftRightTone)                         3.10e+02 5.56e+01      NA      NA
s(measurement.no,wordPos)                    3.31e+02 8.99e+01      NA      NA
s(f0Zscore2,wordPos)                         3.31e+02 5.39e+01      NA      NA
s(durationZscore2,wordPos)                   3.31e+02 2.80e+01      NA      NA
s(measurement.no,wordLeftRightTone)          3.10e+02 1.70e+02      NA      NA
s(f0Zscore2,wordLeftRightTone)               3.10e+02 7.83e+01      NA      NA
s(durationZscore2,wordLeftRightTone)         3.10e+02 3.15e+01      NA      NA
s(measurement.no,word)                       2.09e+02 3.34e+01      NA      NA
s(f0Zscore2,word)                            2.09e+02 5.79e+00      NA      NA
s(durationZscore2,word)                      2.09e+02 1.40e-03      NA      NA
s(measurement.no,speaker)                    1.00e+02 6.41e+01    1.05    1.00
s(durationZscore2,speaker)                   3.00e+02 9.66e+00    1.00    0.45
s(f0Zscore2,speaker)                         1.00e+02 4.03e+01    1.00    0.57
s(measurement.no,speakerPos)                 3.00e+02 4.86e+01    1.05    1.00
s(durationZscore2,speakerPos)                9.00e+02 1.42e+02    1.00    0.46
s(f0Zscore2,speakerPos)                      3.00e+02 6.53e+01    1.00    0.55
s(measurement.no,speakerLeftRightTone)       5.90e+02 1.13e+02    1.05    1.00
s(durationZscore2,speakerLeftRightTone)      1.77e+03 2.52e+02    1.00    0.49
s(f0Zscore2,speakerLeftRightTone)            5.90e+02 5.76e+01    1.00    0.52

# absolute value of durations at 25%, 50% and 75%
duration0.15 = quantile(data.ai.mas$durationZscore2, c(0.15)) * global_sdd + global_meand

duration0.25 = quantile(data.ai.mas$durationZscore2, c(0.25)) * global_sdd + global_meand

duration0.5 = quantile(data.ai.mas$durationZscore2, c(0.5)) * global_sdd + global_meand

duration0.75 = quantile(data.ai.mas$durationZscore2, c(0.75)) * global_sdd + global_meand

duration0.85 = quantile(data.ai.mas$durationZscore2, c(0.85)) * global_sdd + global_meand

# Plotting
# 3D plot 
# png("pred2-1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
# dev.off()
# png("pred2-2.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
# dev.off()
# png("pred2-3.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# dev.off()

# Plotting
# 3D plot 
 png("pred2-1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, nthreads = ncores, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
 dev.off()

null device 
          1 

 png("pred2-2.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
 dev.off()

null device 
          1 

 png("pred2-3.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, 
        zlim = c(450,750),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(450,750), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
 dev.off()

null device 
          1 

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2a, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.ai.mas$durationZscore2, c(0.1,0.9), na.rm = T),
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sdd + global_meand), 
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           alpha.diff = 0.5,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 30 values ranging from -0.786628 to 1.340871. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

abline(h = (1.2 * global_sdd + global_meand), lty=2,lwd=2, col = "white")

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff(gamm.model2a, view="measurement.no", comp=list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8,0.2), na.rm = T)), cond = list(durationZscore2 = 1.3), 
          rm.ranef=TRUE, shade = F, main = "", xlab = "Time (normalized)", ylab = "Difference in F1 (Z)", xaxt = "n", font.lab = 2, cex.lab = 1, cex.axis = 1, lwd = 2, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; set to the value(s): 1.3. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

Difference is not significant.

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

#png("diffduration1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model2a, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.85))), col = "black", lwd = 4, xlab = "Time (Normalized)", ylab = "F1 (Z)", font.lab = 2, xaxt = "n", hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 1.06834268357208. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

plot_smooth(gamm.model2a, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))), col = "royalblue4", add = T, lwd = 4)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))), col = "chartreuse4", add = T, lwd = 4)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))), col = "orange", add = T, lwd = 4)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.15))), col = "red", add = T, lwd = 4)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.644619456704238. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

legend("topright", legend=c("15%", "25%", "50%", "75%", "85%"), 
       col=c("red","orange", "chartreuse4", "royalblue4", "black"), lwd=4)

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

#dev.off()

#png("diffduration2.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_smooth(gamm.model2a, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.85))), col = "black", lwd = 4, xlab = "Time (ms)", ylab = "F1 (Hz)", font.lab = 2, transform.view = function(measurement.no) measurement.no * duration0.85 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1, xaxt = "n", hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 1.06834268357208. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model2a, view="measurement.no",
             rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))), col = "royalblue4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * duration0.75 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))), col = "chartreuse4", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * duration0.5 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))), col = "orange", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * duration0.25 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: X-values are transformed.

plot_smooth(gamm.model2a, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.5), na.rm=T), durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.15))), col = "red", add = T, lwd = 4, transform.view = function(measurement.no) measurement.no * duration0.15 * 0.1, transform = function(f1Zscore2) f1Zscore2 * global_sd1 + global_mean1)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; set to the value(s): -0.0524849620085245. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.644619456704238. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: X-values are transformed.

legend("topright", legend=c("15%", "25%", "50%", "75%", "85%"), 
       col=c("red","orange", "chartreuse4", "royalblue4", "black"), lwd=4)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

#dev.off()

# Generate the high resolution image for the paper
# output combined plot of 1000 dpi
png("pred2-diff-1c.png", width = 9, height = 7.5, units = "in", res = 1000)

layout(matrix(c(1, 2), nrow = 2, byrow = TRUE), heights = c(2, 1))

par(oma = c(4, 0, 1, 0), xaxs = "i", yaxs = "i")  

x_range <- range(data.ai.mas$measurement.no, na.rm = TRUE)


par(mar = c(0.5, 4, 1.5, 2)) 
plot_diff2(gamm.model2a,
           view = c("measurement.no", "durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8, 0.2), na.rm = TRUE)),
           ylim = quantile(data.ai.mas$durationZscore2, c(0.05, 0.95), na.rm = TRUE),
           xlim = x_range,
           n.grid = 100,
           transform.view = c(function(measurement.no) measurement.no,
                              function(durationZscore2) durationZscore2 * global_sdd + global_meand), 
           sim.ci = FALSE,
           show.diff = FALSE,
           add.color.legend = FALSE,
           rm.ranef = TRUE,  
           xlab = "", ylab = "Duration (ms)",
           font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n",
           color = mapcols_pastel, hide.label = TRUE)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 100 values ranging from -1.043110 to 1.889298. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

abline(h = (1.2 * global_sdd + global_meand), lty = 2, lwd = 2, col = "white")

par(mar = c(2, 4, 0.5, 2))
plot_diff(gamm.model2a,
          view = "measurement.no",
          comp = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8, 0.2), na.rm = TRUE)),
          cond = list(durationZscore2 = 1.2),
          rm.ranef = TRUE,
          shade = FALSE,
          main = "",
          xlab = "", ylab = "Diff in F1 (Z)",
          xlim = x_range,
          xaxt = "n",
          font.lab = 2, cex.lab = 1, cex.axis = 1, lwd = 2, hide.label = TRUE)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; set to the value(s): 1.2. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

Difference is not significant.

axis(1, at = tickvals, labels = ticknames, las = 1, cex.axis = 1)

mtext("Time (Normalized)", side = 1, outer = TRUE, line = 2.5, font = 2, cex = 1.3)

dev.off()

null device 
          1 

layout(matrix(c(1, 2), nrow = 2, byrow = TRUE), heights = c(2, 1))


par(oma = c(4, 0, 1, 0), xaxs = "i", yaxs = "i")

x_range <- range(data.ai.mas$measurement.no, na.rm = TRUE)


par(mar = c(0.5, 4, 1.5, 2))
plot_diff2(gamm.model2a,
           view = c("measurement.no", "durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8, 0.2), na.rm = TRUE)),
           ylim = quantile(data.ai.mas$durationZscore2, c(0.05, 0.95), na.rm = TRUE),
           xlim = x_range,
           transform.view = c(function(measurement.no) measurement.no,
                              function(durationZscore2) durationZscore2 * global_sdd + global_meand), 
           sim.ci = FALSE,
           show.diff = FALSE,
           add.color.legend = FALSE,
           rm.ranef = TRUE,  
           xlab = "", ylab = "Duration (ms)",
           font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n",
           color = mapcols_pastel, hide.label = TRUE)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 30 values ranging from -1.043110 to 1.889298. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

abline(h = (1.2 * global_sdd + global_meand), lty = 2, lwd = 2, col = "white")

par(mar = c(2, 4, 0.5, 2))
plot_diff(gamm.model2a,
          view = "measurement.no",
          comp = list(f0Zscore2 = quantile(data.ai.mas$f0Zscore2, c(0.8, 0.2), na.rm = TRUE)),
          cond = list(durationZscore2 = 1.2),
          rm.ranef = TRUE,
          shade = FALSE,
          main = "",
          xlab = "", ylab = "Diff in F1 (Z)",
          xlim = x_range,
          xaxt = "n",
          font.lab = 2, cex.lab = 1, cex.axis = 1, lwd = 2, hide.label = TRUE)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; set to the value(s): 1.2. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

Difference is not significant.

axis(1, at = tickvals, labels = ticknames, las = 1, cex.axis = 1)


mtext("Time (Normalized)", side = 1, outer = TRUE, line = 2.5, font = 2, cex = 1.3)

5.2.1 Composition of f0 contours of different tones and the output of the second model.

# Modelling F2 ~ tone, with no autocorrelation

system.time(gamm.f0model.noAR <- bam(f0Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr"),
                        
                        
                        
                        
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

utilisateur     système      écoulé 
      0.479       0.012       0.124 

r.gamm.f0model <- start_value_rho(gamm.f0model.noAR)

# Modelling F2 ~ tone

# Final model with auto-correlation

system.time(gamm.f0model <- bam(f0Zscore2 ~ toneBis.ord + 
                        # smooth
                        s(measurement.no, bs="cr") +
                          
                        # smooth by factors
                        s(measurement.no, by=toneBis.ord, bs="cr"),
                        
                        
                        
                        
                    
                      data=data.ai.mas, method="fREML", rho = r.gamm.f0model, AR.start = data.ai.mas$start, discrete = TRUE, nthreads = ncores)
)

utilisateur     système      écoulé 
      0.287       0.005       0.042 

#Plotting
### retested ordered factors 

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.05,0.95), na.rm = T),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (Normalized)", ylab = "F0 (Z)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = "bw", xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.547736 to 1.651185. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

plot_smooth(gamm.f0model, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = quantile(data.ai.mas$f0Zscore2, c(0.05,0.95), na.rm = T),,
            lwd = 4, xlab = "", ylab = "", xaxt = "n", font.lab = 2, add = T)

Summary:
    * toneBis.ord : factor; set to the value(s): 1. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * NOTE : No random effects in the model to cancel.
 

plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), 
            col = "orange", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 2. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * NOTE : No random effects in the model to cancel.
 

plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), 
            col = "chartreuse4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 3. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * NOTE : No random effects in the model to cancel.
 

plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), 
            col = "royalblue4", add = T, lwd = 4)

Summary:
    * toneBis.ord : factor; set to the value(s): 4. 
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * NOTE : No random effects in the model to cancel.
 

legend("bottomright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=2)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

NA
NA
NA

#Plotting
### retested ordered factors
png("composition1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2a, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.05,0.95), na.rm = T),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (Normalized)", ylab = "F0 (Z)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = "bw", xaxt = "n")
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)




plot_smooth(gamm.f0model, view="measurement.no",
            main = "", rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "1"), col = "red", ylim = quantile(data.ai.mas$f0Zscore2, c(0.05,0.95), na.rm = T),,
            lwd = 4, xlab = "", ylab = "", xaxt = "n", font.lab = 2, add = T)
plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "2"), 
            col = "orange", add = T, lwd = 4)
plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "3"), 
            col = "chartreuse4", add = T, lwd = 4)
plot_smooth(gamm.f0model, view="measurement.no",
            rug=F, rm.ranef = T, shade = F, cond = list(toneBis.ord = "4"), 
            col = "royalblue4", add = T, lwd = 4)
legend("bottomright", legend=c("Tone 1", "Tone 2", "Tone 3", "Tone 4"), 
       col=c("red","orange", "chartreuse4", "royalblue4"), lwd=2)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

dev.off()

5.3 For Appendix C: test the model with te-tensor

system.time(gamm.model2ate.noAR <- bam(f1Zscore2 ~ 
                        # te tensor
                        
                        te(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2ate.noAR, paste("Gamm_model2ate_noAR.rds"))

gamm.model2ate.noAR <- 
  readRDS("Gamm_model2ate_noAR.rds")

r.gamm.model2ate <- start_value_rho(gamm.model2ate.noAR)

system.time(gamm.model2ate <- bam(f1Zscore2 ~ 
                        # te tensor
                        
                        te(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
                      
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2ate, AR.start = data.ai.mas$start)
)

#saveRDS(gamm.model2ate, paste("Gamm_model2ate.rds"))

gamm.model2ate <- 
  readRDS("Gamm_model2ate.rds")

summary(gamm.model2ate, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ te(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
    k = c(5, 8, 12)) + s(word, bs = "re") + s(wordPos, bs = "re") + 
    s(wordLeftRightTone, bs = "re") + s(wordPos, measurement.no, 
    bs = "re") + s(wordPos, f0Zscore2, bs = "re") + s(wordPos, 
    durationZscore2, bs = "re") + s(wordLeftRightTone, measurement.no, 
    bs = "re") + s(wordLeftRightTone, f0Zscore2, bs = "re") + 
    s(wordLeftRightTone, durationZscore2, bs = "re") + s(word, 
    measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.05259    0.05057   -1.04    0.298

Approximate significance of smooth terms:
                                               edf Ref.df     F p-value    
te(measurement.no,f0Zscore2,durationZscore2) 54.32  71.92 13.15  <2e-16 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.771   Deviance explained = 80.3%
fREML = 6009.3  Scale est. = 0.13463   n = 11029

gam.check(gamm.model2ate)

Method: fREML   Optimizer: perf chol
$grad
 [1]  7.212009e-13 -2.522427e-13 -1.788791e-12  3.304024e-12 -1.247003e-12 -1.829648e-12
 [7]  4.263256e-14  8.171241e-14 -5.151435e-14  6.394885e-13 -3.197442e-13 -3.215206e-13
[13] -6.288303e-13  3.552714e-15 -9.134682e-05 -9.023893e-13 -9.264209e-05 -1.367795e-13
[19] -9.264209e-05 -4.014566e-13 -9.264209e-05  1.172396e-13  5.173639e-14  0.000000e+00
[25]  1.061373e-13  1.776357e-14  4.130030e-14 -1.563194e-13  3.774758e-14  3.979039e-13
[31]  3.730349e-14 -1.385558e-13 -8.393286e-14  1.455192e-11

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]
   5.188001e+00  7.070166e-02  1.115972e-01 -1.147504e-02  9.895278e-03  3.525500e-03
   7.070166e-02  9.012753e-01  4.572425e-01  8.988453e-02  2.261362e-02  3.699769e-02
   1.115972e-01  4.572425e-01  1.850239e+00  1.608275e-01  5.109433e-02  2.520025e-02
  -1.147504e-02  8.988453e-02  1.608275e-01  1.946286e+01  4.299996e+00  3.830146e+00
   9.895278e-03  2.261362e-02  5.109433e-02  4.299996e+00  9.064811e+00  7.070647e-01
   3.525500e-03  3.699769e-02  2.520025e-02  3.830146e+00  7.070647e-01  7.531835e+00
   1.396767e-02  1.015448e-01  9.147414e-02 -7.443352e-01  6.573249e-01  1.191398e-01
  -1.567791e-03  5.247568e-02  6.785587e-03 -7.783780e-02  4.203683e-01  1.004370e-01
  -2.376092e-03 -1.975131e-02  1.076698e-03  8.104132e-01  2.859538e-01  3.205752e-01
   4.822148e-04  1.540589e-01  2.596237e-01 -3.135857e+00 -6.797270e-01 -2.317508e+00
   1.953989e-02  1.173771e-01  1.956216e-01  1.129581e-02 -8.133214e-02 -3.831317e-01
   9.154741e-03 -3.501889e-02 -6.379211e-02  5.933564e-01  2.865775e-01  6.041066e-01
   7.658543e-04  2.644431e-02  7.693600e-02 -1.316617e+00 -2.380616e-01 -2.725343e-01
   2.047425e-03  3.872940e-03  9.241259e-03  8.669250e-03  5.948245e-03  1.468515e-02
   6.492905e-07  7.584493e-07  1.101168e-06  5.555260e-05  2.305081e-05  2.587461e-05
   8.540463e-02  1.604533e-02  6.317508e-02 -7.677012e-02 -3.605730e-02 -2.156512e-02
   9.413137e-07  2.165759e-06  1.267550e-05 -1.246160e-06  4.798679e-06 -2.154672e-05
   7.555759e-04  2.421823e-03  8.521740e-03  1.031988e-01  1.876452e-02  3.568459e-02
   9.413137e-07  2.165759e-06  1.267550e-05 -1.246160e-06  4.798679e-06 -2.154672e-05
   1.558590e-02 -1.113138e-01  1.182997e-01  4.005980e-02 -1.391299e-02 -4.520616e-02
   9.413138e-07  2.165759e-06  1.267550e-05 -1.246160e-06  4.798679e-06 -2.154672e-05
   1.103605e-02 -2.607238e-02 -1.440185e-02  3.551112e-01  1.816060e-01  5.029666e-02
   1.204923e-03 -3.165195e-03 -1.101602e-02 -4.121962e-02 -7.968617e-03 -1.805989e-02
   1.112656e-02  3.392380e-02  2.834617e-02  5.944516e-01  3.532166e-01  5.319058e-01
   1.204923e-03 -3.165195e-03 -1.101602e-02 -4.121962e-02 -7.968617e-03 -1.805989e-02
   1.057317e-02 -2.812451e-02 -1.271693e-02  1.915228e-01 -2.517857e-01  7.876653e-02
   1.204923e-03 -3.165195e-03 -1.101602e-02 -4.121962e-02 -7.968616e-03 -1.805989e-02
   4.483139e-02  1.891545e-01  1.770829e-02  1.195657e-01 -5.655524e-02  2.430439e-01
   7.039883e-04  1.909300e-03  8.748746e-03  1.703776e-02  7.875045e-03 -6.890274e-02
           [,7]          [,8]          [,9]         [,10]         [,11]         [,12]
   1.396767e-02 -1.567791e-03 -2.376092e-03  4.822148e-04  1.953989e-02  9.154741e-03
   1.015448e-01  5.247568e-02 -1.975131e-02  1.540589e-01  1.173771e-01 -3.501889e-02
   9.147414e-02  6.785587e-03  1.076698e-03  2.596237e-01  1.956216e-01 -6.379211e-02
  -7.443352e-01 -7.783780e-02  8.104132e-01 -3.135857e+00  1.129581e-02  5.933564e-01
   6.573249e-01  4.203683e-01  2.859538e-01 -6.797270e-01 -8.133214e-02  2.865775e-01
   1.191398e-01  1.004370e-01  3.205752e-01 -2.317508e+00 -3.831317e-01  6.041066e-01
   1.484122e+01  4.635109e-01 -3.397288e-01  9.540342e+00  9.281281e-02 -3.065428e-01
   4.635109e-01  7.541022e+00  1.351556e-01  1.943293e-01  3.270023e+00 -2.878651e-02
  -3.397288e-01  1.351556e-01  3.584680e+00 -4.106119e-01 -2.416990e-01  1.726895e+00
   9.540342e+00  1.943293e-01 -4.106119e-01  4.315220e+01 -2.123477e-01 -3.864110e-01
   9.281281e-02  3.270023e+00 -2.416990e-01 -2.123477e-01  7.852431e+00 -3.200757e-01
  -3.065428e-01 -2.878651e-02  1.726895e+00 -3.864110e-01 -3.200757e-01  3.268123e+00
   2.912164e+00 -2.732735e-02 -1.033708e-01  7.687327e+00  1.876938e-01 -1.014930e-01
   1.550674e-02  4.179317e-01  1.986650e-02  3.003052e-02  4.907111e-01 -1.934225e-02
  -2.327439e-05  1.196505e-05  1.805820e-04 -3.343304e-05 -2.267541e-05  1.928256e-04
   8.098802e-02  2.443496e-02 -3.570246e-02  1.685606e-01  1.225256e-01  3.187835e-02
   2.859237e-06 -1.337164e-05  3.076987e-06 -7.991972e-07 -6.619546e-06 -1.629600e-05
  -4.098497e-03 -8.939429e-03 -1.943172e-02 -5.706676e-02 -1.181631e-02 -1.125223e-02
   2.859237e-06 -1.337164e-05  3.076987e-06 -7.991972e-07 -6.619546e-06 -1.629600e-05
  -4.444425e-02  5.093016e-02 -4.151662e-02 -8.489854e-02 -2.481049e-01  1.705532e-02
   2.859237e-06 -1.337164e-05  3.076987e-06 -7.991972e-07 -6.619546e-06 -1.629600e-05
  -1.443533e-01 -8.831735e-04 -5.278766e-02 -3.330908e-01  1.061155e-01 -4.849169e-02
  -2.724834e-03 -4.684749e-02 -3.789710e-02  6.493868e-03 -3.854051e-02 -7.794985e-03
  -2.741598e-01 -3.962282e-01  1.638019e-01 -8.334225e-01  1.355954e-01  1.509934e-01
  -2.724834e-03 -4.684749e-02 -3.789710e-02  6.493868e-03 -3.854051e-02 -7.794985e-03
  -2.159544e-01 -4.495118e-01  4.878371e-02 -3.881891e-01  7.156239e-02  1.840412e-01
  -2.724834e-03 -4.684749e-02 -3.789710e-02  6.493868e-03 -3.854051e-02 -7.794985e-03
   7.704296e-01  3.334790e-01  3.799677e-02  2.669121e+00  9.832643e-01  1.042509e-02
  -2.335171e-02 -3.912630e-04 -7.549618e-02  8.387261e-02 -1.839649e-02 -4.976510e-02
          [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   7.658543e-04  2.047425e-03  6.492905e-07  8.540463e-02  9.413137e-07  7.555759e-04
   2.644431e-02  3.872940e-03  7.584493e-07  1.604533e-02  2.165759e-06  2.421823e-03
   7.693600e-02  9.241259e-03  1.101168e-06  6.317508e-02  1.267550e-05  8.521740e-03
  -1.316617e+00  8.669250e-03  5.555260e-05 -7.677012e-02 -1.246160e-06  1.031988e-01
  -2.380616e-01  5.948245e-03  2.305081e-05 -3.605730e-02  4.798679e-06  1.876452e-02
  -2.725343e-01  1.468515e-02  2.587461e-05 -2.156512e-02 -2.154672e-05  3.568459e-02
   2.912164e+00  1.550674e-02 -2.327439e-05  8.098802e-02  2.859237e-06 -4.098497e-03
  -2.732735e-02  4.179317e-01  1.196505e-05  2.443496e-02 -1.337164e-05 -8.939429e-03
  -1.033708e-01  1.986650e-02  1.805820e-04 -3.570246e-02  3.076987e-06 -1.943172e-02
   7.687327e+00  3.003052e-02 -3.343304e-05  1.685606e-01 -7.991972e-07 -5.706676e-02
   1.876938e-01  4.907111e-01 -2.267541e-05  1.225256e-01 -6.619546e-06 -1.181631e-02
  -1.014930e-01 -1.934225e-02  1.928256e-04  3.187835e-02 -1.629600e-05 -1.125223e-02
   3.381713e+00 -1.118427e-02 -8.776317e-06  6.630582e-02  9.706401e-07 -6.710163e-03
  -1.118427e-02  1.015506e-01 -1.476555e-06  1.794625e-02 -2.181901e-07  1.759602e-03
  -8.776317e-06 -1.476555e-06  9.136783e-05  2.062868e-06 -2.992420e-09  1.532424e-06
   6.630582e-02  1.794625e-02  2.062868e-06  2.943206e+01  8.965843e-06 -2.558432e-03
   9.706401e-07 -2.181901e-07 -2.992420e-09  8.965843e-06  9.270400e-05  5.194029e-06
  -6.710163e-03  1.759602e-03  1.532424e-06 -2.558432e-03  5.194029e-06  1.457989e-01
   9.706401e-07 -2.181901e-07 -2.992420e-09  8.965843e-06  6.191423e-08  5.194029e-06
  -1.839253e-02  2.297464e-03 -1.960974e-06  8.176173e-03 -9.523079e-05 -2.141524e-03
   9.706401e-07 -2.181901e-07 -2.992420e-09  8.965843e-06  6.191423e-08  5.194029e-06
  -1.364824e-02  1.713675e-02 -2.616396e-06  9.186364e-01  3.469226e-06  5.542084e-03
  -4.306343e-03 -4.453118e-03 -3.942696e-07  5.025075e-03  5.194090e-05 -4.597078e-04
  -2.349707e-01 -1.118681e-02 -1.523626e-05 -1.879890e-02  2.174704e-05  7.331150e-01
  -4.306343e-03 -4.453118e-03 -3.942696e-07  5.025075e-03  5.194090e-05 -4.597078e-04
  -1.223800e-02  1.402012e-02  1.297995e-05  1.316466e-02 -3.523722e-05  8.839315e-03
  -4.306343e-03 -4.453118e-03 -3.942696e-07  5.025075e-03  5.194090e-05 -4.597078e-04
   5.623135e-01  1.277940e-02  5.329717e-07  7.341681e-01  4.296872e-06  1.091782e-02
   1.005554e-02 -7.064315e-03 -6.826775e-06 -7.623853e-04  9.606364e-06  3.571841e-04
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]
   9.413137e-07  1.558590e-02  9.413138e-07  1.103605e-02  1.204923e-03  1.112656e-02
   2.165759e-06 -1.113138e-01  2.165759e-06 -2.607238e-02 -3.165195e-03  3.392380e-02
   1.267550e-05  1.182997e-01  1.267550e-05 -1.440185e-02 -1.101602e-02  2.834617e-02
  -1.246160e-06  4.005980e-02 -1.246160e-06  3.551112e-01 -4.121962e-02  5.944516e-01
   4.798679e-06 -1.391299e-02  4.798679e-06  1.816060e-01 -7.968617e-03  3.532166e-01
  -2.154672e-05 -4.520616e-02 -2.154672e-05  5.029666e-02 -1.805989e-02  5.319058e-01
   2.859237e-06 -4.444425e-02  2.859237e-06 -1.443533e-01 -2.724834e-03 -2.741598e-01
  -1.337164e-05  5.093016e-02 -1.337164e-05 -8.831735e-04 -4.684749e-02 -3.962282e-01
   3.076987e-06 -4.151662e-02  3.076987e-06 -5.278766e-02 -3.789710e-02  1.638019e-01
  -7.991972e-07 -8.489854e-02 -7.991972e-07 -3.330908e-01  6.493868e-03 -8.334225e-01
  -6.619546e-06 -2.481049e-01 -6.619546e-06  1.061155e-01 -3.854051e-02  1.355954e-01
  -1.629600e-05  1.705532e-02 -1.629600e-05 -4.849169e-02 -7.794985e-03  1.509934e-01
   9.706401e-07 -1.839253e-02  9.706401e-07 -1.364824e-02 -4.306343e-03 -2.349707e-01
  -2.181901e-07  2.297464e-03 -2.181901e-07  1.713675e-02 -4.453118e-03 -1.118681e-02
  -2.992420e-09 -1.960974e-06 -2.992420e-09 -2.616396e-06 -3.942696e-07 -1.523626e-05
   8.965843e-06  8.176173e-03  8.965843e-06  9.186364e-01  5.025075e-03 -1.879890e-02
   6.191423e-08 -9.523079e-05  6.191423e-08  3.469226e-06  5.194090e-05  2.174704e-05
   5.194029e-06 -2.141524e-03  5.194029e-06  5.542084e-03 -4.597078e-04  7.331150e-01
   9.270400e-05 -9.523079e-05  6.191423e-08  3.469226e-06  5.194090e-05  2.174704e-05
  -9.523079e-05  1.103906e+01 -9.523079e-05  6.351649e-03 -6.777299e-02  2.260902e-02
   6.191423e-08 -9.523079e-05  9.270400e-05  3.469226e-06  5.194090e-05  2.174704e-05
   3.469226e-06  6.351649e-03  3.469226e-06  7.779513e+00 -3.974299e-02 -3.554961e-02
   5.194090e-05 -6.777299e-02  5.194090e-05 -3.974299e-02  2.129867e-01  2.935215e-02
   2.174704e-05  2.260902e-02  2.174704e-05 -3.554961e-02  2.935215e-02  1.822489e+01
   5.194090e-05 -6.777299e-02  5.194090e-05 -3.974299e-02  2.129867e-01  2.935215e-02
  -3.523722e-05  2.680778e+00 -3.523722e-05 -4.409252e-02 -6.687288e-03 -2.346036e-01
   5.194090e-05 -6.777299e-02  5.194090e-05 -3.974299e-02  2.129867e-01  2.935215e-02
   4.296872e-06 -3.474327e-02  4.296872e-06  7.237061e-01  1.297742e-02  1.392221e-01
   9.606364e-06 -5.762978e-02  9.606364e-06 -4.462874e-03  1.514256e-02  7.467411e-03
          [,25]         [,26]         [,27]         [,28]         [,29]         [,30]
   1.204923e-03  1.057317e-02  1.204923e-03  4.483139e-02  7.039883e-04 -2.245349e-03
  -3.165195e-03 -2.812451e-02 -3.165195e-03  1.891545e-01  1.909300e-03  1.386755e-01
  -1.101602e-02 -1.271693e-02 -1.101602e-02  1.770829e-02  8.748746e-03  1.835951e-01
  -4.121962e-02  1.915228e-01 -4.121962e-02  1.195657e-01  1.703776e-02  1.412183e+00
  -7.968617e-03 -2.517857e-01 -7.968616e-03 -5.655524e-02  7.875045e-03  7.290499e-01
  -1.805989e-02  7.876653e-02 -1.805989e-02  2.430439e-01 -6.890274e-02  4.009883e-01
  -2.724834e-03 -2.159544e-01 -2.724834e-03  7.704296e-01 -2.335171e-02 -2.604793e-01
  -4.684749e-02 -4.495118e-01 -4.684749e-02  3.334790e-01 -3.912630e-04 -2.387759e-01
  -3.789710e-02  4.878371e-02 -3.789710e-02  3.799677e-02 -7.549618e-02  4.678023e-01
   6.493868e-03 -3.881891e-01  6.493868e-03  2.669121e+00  8.387261e-02 -9.756334e-01
  -3.854051e-02  7.156239e-02 -3.854051e-02  9.832643e-01 -1.839649e-02 -3.102992e-01
  -7.794985e-03  1.840412e-01 -7.794985e-03  1.042509e-02 -4.976510e-02  4.793494e-01
  -4.306343e-03 -1.223800e-02 -4.306343e-03  5.623135e-01  1.005554e-02 -8.571114e-02
  -4.453118e-03  1.402012e-02 -4.453118e-03  1.277940e-02 -7.064315e-03 -1.804930e-02
  -3.942696e-07  1.297995e-05 -3.942696e-07  5.329717e-07 -6.826775e-06 -2.735796e-05
   5.025075e-03  1.316466e-02  5.025075e-03  7.341681e-01 -7.623853e-04  8.896215e-02
   5.194090e-05 -3.523722e-05  5.194090e-05  4.296872e-06  9.606364e-06  1.318269e-04
  -4.597078e-04  8.839315e-03 -4.597078e-04  1.091782e-02  3.571841e-04  1.035238e+00
   5.194090e-05 -3.523722e-05  5.194090e-05  4.296872e-06  9.606364e-06  1.318269e-04
  -6.777299e-02  2.680778e+00 -6.777299e-02 -3.474327e-02 -5.762978e-02 -2.943663e-01
   5.194090e-05 -3.523722e-05  5.194090e-05  4.296872e-06  9.606364e-06  1.318269e-04
  -3.974299e-02 -4.409252e-02 -3.974299e-02  7.237061e-01 -4.462874e-03  7.064571e-02
   2.129867e-01 -6.687288e-03  2.129867e-01  1.297742e-02  1.514256e-02  3.372954e-02
   2.935215e-02 -2.346036e-01  2.935215e-02  1.392221e-01  7.467411e-03  8.423757e+00
   2.129867e-01 -6.687288e-03  2.129867e-01  1.297742e-02  1.514256e-02  3.372954e-02
  -6.687288e-03  1.187852e+01 -6.687288e-03  2.845516e-01  2.043815e-02  9.474120e-02
   2.129867e-01 -6.687288e-03  2.129867e-01  1.297742e-02  1.514256e-02  3.372954e-02
   1.297742e-02  2.845516e-01  1.297742e-02  1.702616e+01 -2.344143e-02 -1.507471e-01
   1.514256e-02  2.043815e-02  1.514256e-02 -2.344143e-02  2.329622e-01  4.788876e-01
          [,31]         [,32]         [,33]         [,34]
   7.039883e-04  5.618612e-03  7.039883e-04 -5.985552e+00
   1.909300e-03 -7.261482e-03  1.909300e-03 -7.167102e+00
   8.748746e-03  5.937652e-02  8.748746e-03 -1.050502e+01
   1.703776e-02  9.407675e-02  1.703776e-02 -4.222852e+01
   7.875045e-03 -8.985932e-03  7.875046e-03 -2.926044e+01
  -6.890274e-02 -1.610393e-01 -6.890274e-02 -2.667084e+01
  -2.335171e-02 -8.872845e-02 -2.335171e-02 -4.544298e+01
  -3.912630e-04  1.540029e-01 -3.912630e-04 -2.715213e+01
  -7.549618e-02 -1.206844e-02 -7.549618e-02 -1.386415e+01
   8.387261e-02  2.474264e-02  8.387261e-02 -8.309164e+01
  -1.839649e-02  8.323468e-01 -1.839649e-02 -3.983799e+01
  -4.976510e-02 -9.317467e-02 -4.976510e-02 -1.503665e+01
   1.005554e-02  2.429064e-02  1.005554e-02 -1.850145e+01
  -7.064315e-03  1.597305e-02 -7.064315e-03 -2.548287e+00
  -6.826775e-06 -4.802593e-06 -6.826775e-06 -9.135131e-04
  -7.623853e-04 -1.435675e-02 -7.623853e-04 -3.577030e+01
   9.606364e-06  7.799399e-06  9.606364e-06 -5.427271e-04
   3.571841e-04  3.800697e-02  3.571841e-04 -4.330749e+00
   9.606364e-06  7.799399e-06  9.606364e-06 -5.427271e-04
  -5.762978e-02  7.943718e-01 -5.762978e-02 -2.001912e+01
   9.606364e-06  7.799399e-06  9.606364e-06 -5.427271e-04
  -4.462874e-03  4.622193e-02 -4.462874e-03 -2.222955e+01
   1.514256e-02  2.880743e-02  1.514256e-02 -1.585267e+00
   7.467411e-03  2.429561e-03  7.467411e-03 -6.974299e+01
   1.514256e-02  2.880743e-02  1.514256e-02 -1.585267e+00
   2.043815e-02  5.117225e-01  2.043815e-02 -3.090183e+01
   1.514256e-02  2.880743e-02  1.514256e-02 -1.585267e+00
  -2.344143e-02 -4.951560e-02 -2.344143e-02 -5.565884e+01
   2.329622e-01  5.057901e-02  2.329622e-01 -2.071437e+00
 [ getOption("max.print") est atteint -- 5 lignes omises ]

Model rank =  8830 / 8830 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value
te(measurement.no,f0Zscore2,durationZscore2) 4.79e+02 5.43e+01    1.01    0.83
s(word)                                      2.09e+02 8.45e+01      NA      NA
s(wordPos)                                   3.31e+02 5.85e+01      NA      NA
s(wordLeftRightTone)                         3.10e+02 5.33e+01      NA      NA
s(measurement.no,wordPos)                    3.31e+02 9.09e+01      NA      NA
s(f0Zscore2,wordPos)                         3.31e+02 5.43e+01      NA      NA
s(durationZscore2,wordPos)                   3.31e+02 2.77e+01      NA      NA
s(measurement.no,wordLeftRightTone)          3.10e+02 1.66e+02      NA      NA
s(f0Zscore2,wordLeftRightTone)               3.10e+02 7.97e+01      NA      NA
s(durationZscore2,wordLeftRightTone)         3.10e+02 3.01e+01      NA      NA
s(measurement.no,word)                       2.09e+02 3.70e+01      NA      NA
s(f0Zscore2,word)                            2.09e+02 5.10e+00      NA      NA
s(durationZscore2,word)                      2.09e+02 2.01e-03      NA      NA
s(measurement.no,speaker)                    1.00e+02 7.15e+01    1.00    0.54
s(durationZscore2,speaker)                   3.00e+02 8.66e+00    0.98    0.12
s(f0Zscore2,speaker)                         1.00e+02 4.00e+01    1.01    0.64
s(measurement.no,speakerPos)                 3.00e+02 4.76e+01    1.00    0.45
s(durationZscore2,speakerPos)                9.00e+02 1.43e+02    0.98    0.21
s(f0Zscore2,speakerPos)                      3.00e+02 6.50e+01    1.01    0.64
s(measurement.no,speakerLeftRightTone)       5.90e+02 1.15e+02    1.00    0.51
s(durationZscore2,speakerLeftRightTone)      1.77e+03 2.52e+02    0.98    0.15
s(f0Zscore2,speakerLeftRightTone)            5.90e+02 5.63e+01    1.01    0.72

5.4 Model 2B: /ai/ female, F1 as output

system.time(gamm.model2b.noAR <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2b.noAR, paste("Gamm_model2b_noAR.rds"))

gamm.model2b.noAR <- 
  readRDS("Gamm_model2b_noAR.rds")

r.gamm.model2b <- start_value_rho(gamm.model2b.noAR)

# Auto-regressive model

system.time(gamm.model2b <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
 
                        
                    
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2b, AR.start = data.ai.fem$start)
)

#saveRDS(gamm.model2b, paste("Gamm_model2b.rds"))

gamm.model2b <- 
  readRDS("Gamm_model2b.rds")

summary(gamm.model2b, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)  0.22118    0.05384   4.108 4.02e-05 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                                edf Ref.df       F p-value    
s(measurement.no)                             8.063  8.233 128.643  <2e-16 ***
s(f0Zscore2)                                  3.926  4.392   3.229  0.0101 *  
s(durationZscore2)                            1.000  1.000   0.217  0.6418    
ti(measurement.no,f0Zscore2,durationZscore2) 34.234 49.415   1.167  0.2020    
ti(measurement.no,durationZscore2)           11.876 14.803  10.933  <2e-16 ***
ti(measurement.no,f0Zscore2)                 16.710 19.949   9.018  <2e-16 ***
ti(f0Zscore2,durationZscore2)                 7.521 10.185   0.421  0.9402    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.821   Deviance explained = 84.6%
fREML = 6042.4  Scale est. = 0.15276   n = 10275

gam.check(gamm.model2b)

Method: fREML   Optimizer: perf chol
$grad
 [1] -1.154632e-13  8.200107e-13 -6.545912e-05  6.128431e-14  7.949197e-14  3.286260e-14
 [7] -6.084022e-14  4.307665e-14 -7.975842e-13 -1.813660e-12 -2.731149e-14 -1.620926e-14
[13] -4.547474e-13 -1.350031e-13 -4.440892e-14 -7.815970e-14  2.220446e-13  6.394885e-13
[19]  1.200817e-12 -5.950795e-14  1.776357e-14 -4.263256e-13  1.030287e-13 -4.178096e-05
[25]  4.156675e-13 -4.280935e-05 -8.402220e-05 -4.280935e-05 -2.415845e-13 -4.280935e-05
[31] -1.350031e-13 -4.307665e-14 -3.410605e-13 -4.063416e-14  1.278977e-13 -3.375078e-14
[37]  5.684342e-14 -1.114664e-13 -1.421085e-13  1.110223e-14  2.700062e-13  1.021405e-13
[43] -2.546585e-11

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]
   3.707736e+00  4.256521e-03 -3.439787e-09  2.442660e-03  3.421694e-03 -1.903607e-03
   4.256521e-03  9.189438e-01  3.220334e-07  1.465172e-02 -3.858111e-03 -1.517274e-02
  -3.439787e-09  3.220334e-07  6.545714e-05 -3.760788e-06  9.383611e-07 -3.327170e-06
   2.442660e-03  1.465172e-02 -3.760788e-06  7.610624e-01  1.448069e-01  4.618826e-01
   3.421694e-03 -3.858111e-03  9.383611e-07  1.448069e-01  1.386301e+00  5.020525e-02
  -1.903607e-03 -1.517274e-02 -3.327170e-06  4.618826e-01  5.020525e-02  1.687185e+00
  -7.502210e-04  2.721292e-03 -7.868513e-07 -2.883147e-02  2.594245e-02  2.371818e-02
  -1.601439e-03 -3.969014e-03 -2.210888e-06  1.840424e-02 -9.027530e-03  9.679929e-02
   3.364277e-03 -1.414825e-02 -5.816796e-08  8.912138e-04 -3.089210e-03  3.334126e-02
   1.136789e-02  5.104034e-02 -7.630238e-07  4.056002e-02  5.407520e-02  9.755006e-02
   1.239257e-03  1.670406e-03  1.548772e-06 -5.751456e-02 -1.982868e-01 -9.729562e-02
   1.102703e-03 -1.660704e-02  5.343195e-06 -1.984499e-02 -8.209846e-02 -5.320451e-02
  -1.578653e-02  1.547437e-02  6.338747e-07 -1.036205e-02  3.574669e-02  5.094951e-02
   2.095772e-03 -3.523878e-03 -4.866094e-07 -6.211851e-03  5.345944e-02 -3.233403e-02
  -2.112522e-03  3.748217e-04 -2.005465e-07 -8.753664e-03  3.638299e-03 -1.226126e-03
  -5.267709e-04  5.218394e-02 -1.246325e-06 -5.760342e-02 -1.567993e-01 -8.965645e-02
  -4.230387e-03 -5.615988e-02  2.149096e-06  3.485224e-02 -5.334621e-02 -4.159954e-02
  -1.450772e-03  1.285785e-02  1.927457e-05 -4.754317e-02  8.315131e-02  5.437533e-02
   5.960704e-03  3.276056e-02  6.151179e-06 -1.224078e-01  6.325730e-02 -3.480853e-02
   5.921876e-04 -1.748254e-02  3.849830e-07  1.903181e-02 -7.069135e-03  3.021515e-02
   1.282710e-03 -2.497039e-02  7.961872e-06  1.264334e-03  2.140173e-02  4.725367e-02
   4.188663e-03 -7.086206e-03  1.218690e-06 -7.315618e-02  1.453096e-02 -5.501964e-03
  -9.935313e-03 -1.378086e-01  1.189511e-06  8.069501e-02 -1.085931e-02 -2.907726e-02
           [,7]          [,8]          [,9]         [,10]         [,11]         [,12]
  -7.502210e-04 -1.601439e-03  3.364277e-03  1.136789e-02  1.239257e-03  1.102703e-03
   2.721292e-03 -3.969014e-03 -1.414825e-02  5.104034e-02  1.670406e-03 -1.660704e-02
  -7.868513e-07 -2.210888e-06 -5.816796e-08 -7.630238e-07  1.548772e-06  5.343195e-06
  -2.883147e-02  1.840424e-02  8.912138e-04  4.056002e-02 -5.751456e-02 -1.984499e-02
   2.594245e-02 -9.027530e-03 -3.089210e-03  5.407520e-02 -1.982868e-01 -8.209846e-02
   2.371818e-02  9.679929e-02  3.334126e-02  9.755006e-02 -9.729562e-02 -5.320451e-02
   1.338760e+00  8.540904e-02 -9.741826e-03 -1.873891e-02  1.562958e-02  3.030190e-02
   8.540904e-02  1.657790e+00 -2.379495e-02  5.874106e-03  9.763144e-03  1.417672e-02
  -9.741826e-03 -2.379495e-02  2.008453e+00  4.425884e-01 -1.229939e-02 -3.852315e-03
  -1.873891e-02  5.874106e-03  4.425884e-01  2.196416e+00 -3.460756e-02 -4.563079e-03
   1.562958e-02  9.763144e-03 -1.229939e-02 -3.460756e-02  4.136977e-01 -1.410149e-02
   3.030190e-02  1.417672e-02 -3.852315e-03 -4.563079e-03 -1.410149e-02  3.356852e-01
  -1.008657e-02  3.166873e-02 -4.212262e-02  6.998535e-02 -9.617042e-03 -2.833595e-02
   2.379904e-03  2.467704e-02  6.401218e-04  1.295302e-02  2.584312e-02  4.070835e-02
  -1.551054e-03 -5.584742e-03  3.312064e-03  2.851291e-03  5.081978e-04 -6.603301e-03
  -2.919315e-03 -2.412261e-02  2.464823e-03  1.074017e-02 -3.186518e-02  1.508749e-02
   1.021466e-02  2.163986e-03  3.026618e-02  3.310892e-02  7.243879e-03  1.399706e-02
  -3.143423e-03  7.629090e-03  3.443392e-03 -1.497605e-02  6.563478e-02 -7.166678e-03
   3.261042e-02  8.454598e-02  2.489642e-02  4.223383e-02 -2.612511e-02 -7.654614e-02
   9.385444e-03  1.381304e-03  1.007014e-02  1.751672e-02  1.712842e-03  1.322051e-03
  -8.661230e-03  1.479329e-02 -1.677438e-03 -9.745778e-03  1.917097e-02 -6.132109e-03
  -1.695322e-03  1.051208e-02  1.364478e-02  2.612878e-02  1.769715e-03 -4.895969e-02
   6.231357e-03 -2.271116e-02  4.063509e-02  1.144204e-01  2.065354e-02  2.706421e-02
          [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -1.578653e-02  2.095772e-03 -2.112522e-03 -5.267709e-04 -4.230387e-03 -1.450772e-03
   1.547437e-02 -3.523878e-03  3.748217e-04  5.218394e-02 -5.615988e-02  1.285785e-02
   6.338747e-07 -4.866094e-07 -2.005465e-07 -1.246325e-06  2.149096e-06  1.927457e-05
  -1.036205e-02 -6.211851e-03 -8.753664e-03 -5.760342e-02  3.485224e-02 -4.754317e-02
   3.574669e-02  5.345944e-02  3.638299e-03 -1.567993e-01 -5.334621e-02  8.315131e-02
   5.094951e-02 -3.233403e-02 -1.226126e-03 -8.965645e-02 -4.159954e-02  5.437533e-02
  -1.008657e-02  2.379904e-03 -1.551054e-03 -2.919315e-03  1.021466e-02 -3.143423e-03
   3.166873e-02  2.467704e-02 -5.584742e-03 -2.412261e-02  2.163986e-03  7.629090e-03
  -4.212262e-02  6.401218e-04  3.312064e-03  2.464823e-03  3.026618e-02  3.443392e-03
   6.998535e-02  1.295302e-02  2.851291e-03  1.074017e-02  3.310892e-02 -1.497605e-02
  -9.617042e-03  2.584312e-02  5.081978e-04 -3.186518e-02  7.243879e-03  6.563478e-02
  -2.833595e-02  4.070835e-02 -6.603301e-03  1.508749e-02  1.399706e-02 -7.166678e-03
   3.060133e+01  3.159382e+00  3.883270e-01 -7.558851e-01  5.507709e-01  2.845231e+00
   3.159382e+00  4.066880e+00  9.823880e-02  6.603126e-03 -4.640528e-01  9.035645e-01
   3.883270e-01  9.823880e-02  2.088139e-01  1.192619e-02 -5.773894e-03  1.153831e-01
  -7.558851e-01  6.603126e-03  1.192619e-02  1.082043e+01  4.186514e-01  1.857964e-01
   5.507709e-01 -4.640528e-01 -5.773894e-03  4.186514e-01  2.051420e+00  6.970931e-02
   2.845231e+00  9.035645e-01  1.153831e-01  1.857964e-01  6.970931e-02  9.392829e+00
  -2.448966e+00 -4.105140e-01 -6.281418e-02  4.972201e+00  5.145318e-01 -6.467412e-01
   2.704772e-01  7.513104e-02  1.204274e-01  1.559655e-01  6.218947e-02  1.705923e-01
   5.992586e-01  2.239934e-01  2.165978e-03 -7.742008e-02  8.740747e-02  1.789560e+00
  -2.051214e+00 -1.608512e-01  5.073363e-02  3.838970e+00  1.064557e-01 -2.970552e-01
   1.736339e+00  4.409520e-02  1.556678e-02  3.067167e-01  7.150232e-01  1.676066e-01
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]
   5.960704e-03  5.921876e-04  1.282710e-03  4.188663e-03 -9.935313e-03 -2.927822e-09
   3.276056e-02 -1.748254e-02 -2.497039e-02 -7.086206e-03 -1.378086e-01  2.069708e-06
   6.151179e-06  3.849830e-07  7.961872e-06  1.218690e-06  1.189511e-06  8.542022e-10
  -1.224078e-01  1.903181e-02  1.264334e-03 -7.315618e-02  8.069501e-02 -4.055333e-06
   6.325730e-02 -7.069135e-03  2.140173e-02  1.453096e-02 -1.085931e-02  3.638271e-06
  -3.480853e-02  3.021515e-02  4.725367e-02 -5.501964e-03 -2.907726e-02  4.895228e-06
   3.261042e-02  9.385444e-03 -8.661230e-03 -1.695322e-03  6.231357e-03 -7.945661e-07
   8.454598e-02  1.381304e-03  1.479329e-02  1.051208e-02 -2.271116e-02  6.371346e-07
   2.489642e-02  1.007014e-02 -1.677438e-03  1.364478e-02  4.063509e-02  6.680684e-07
   4.223383e-02  1.751672e-02 -9.745778e-03  2.612878e-02  1.144204e-01 -1.998127e-08
  -2.612511e-02  1.712842e-03  1.917097e-02  1.769715e-03  2.065354e-02  1.141493e-06
  -7.654614e-02  1.322051e-03 -6.132109e-03 -4.895969e-02  2.706421e-02 -4.772511e-06
  -2.448966e+00  2.704772e-01  5.992586e-01 -2.051214e+00  1.736339e+00  1.836389e-04
  -4.105140e-01  7.513104e-02  2.239934e-01 -1.608512e-01  4.409520e-02  4.122814e-05
  -6.281418e-02  1.204274e-01  2.165978e-03  5.073363e-02  1.556678e-02  2.582221e-06
   4.972201e+00  1.559655e-01 -7.742008e-02  3.838970e+00  3.067167e-01  7.603453e-06
   5.145318e-01  6.218947e-02  8.740747e-02  1.064557e-01  7.150232e-01  1.234079e-05
  -6.467412e-01  1.705923e-01  1.789560e+00 -2.970552e-01  1.676066e-01  3.699203e-04
   2.381442e+01  2.875230e-01 -1.373224e-01  6.895438e+00  7.449990e-01 -3.816212e-05
   2.875230e-01  7.154664e-01  3.946369e-02  9.198456e-02  3.303410e-01  1.402755e-05
  -1.373224e-01  3.946369e-02  1.422359e+00 -6.214697e-02  1.689112e-01  1.378430e-04
   6.895438e+00  9.198456e-02 -6.214697e-02  9.064746e+00  9.947550e-02 -3.106739e-05
   7.449990e-01  3.303410e-01  1.689112e-01  9.947550e-02  6.282353e+00  3.348708e-05
          [,25]         [,26]         [,27]         [,28]         [,29]         [,30]
  -1.126736e-01  2.063706e-07  5.679475e-10  2.063706e-07  3.882457e-03  2.063706e-07
  -1.599335e-02 -1.424205e-06 -2.295523e-07 -1.424205e-06  1.981334e-01 -1.424205e-06
   6.283127e-07  1.000627e-10 -2.313943e-10  1.000627e-10 -4.814658e-07  1.000627e-10
   1.359483e-02  2.833859e-07 -1.032704e-07  2.833859e-07 -9.504335e-03  2.833859e-07
   6.263286e-02  2.406207e-07  1.117282e-07  2.406207e-07  3.514388e-03  2.406207e-07
  -1.638895e-02  1.265264e-06  3.374885e-08  1.265264e-06 -4.315125e-02  1.265264e-06
  -1.022864e-02 -4.546403e-07  2.608783e-08 -4.546403e-07 -2.099338e-02 -4.546403e-07
  -3.968230e-02 -4.601972e-07  4.017057e-08 -4.601972e-07  4.877339e-03 -4.601972e-07
  -1.169641e-02 -1.162130e-07  9.070731e-08 -1.162130e-07 -9.552986e-03 -1.162130e-07
  -5.077127e-02  1.420412e-06  1.903548e-07  1.420412e-06  4.244325e-02  1.420412e-06
  -1.991763e-02  1.042390e-06  1.714204e-07  1.042390e-06 -5.752209e-02  1.042390e-06
   1.164458e-02  5.515525e-07  2.420428e-07  5.515525e-07  5.603375e-03  5.515525e-07
  -1.754743e-01  7.199164e-06  8.696583e-07  7.199164e-06  9.546758e-02  7.199164e-06
  -4.420270e-02 -3.065338e-06  2.447069e-06 -3.065338e-06  2.304229e-02 -3.065338e-06
   2.382805e-03 -7.383779e-07 -1.430373e-07 -7.383779e-07  3.352754e-03 -7.383779e-07
   3.544592e-02 -1.411772e-06  1.244612e-06 -1.411772e-06  2.575157e-02 -1.411772e-06
   8.010500e-03  3.089144e-06 -2.148769e-07  3.089144e-06 -2.834340e-02  3.089144e-06
   8.395095e-04  6.285854e-06  2.143050e-07  6.285854e-06 -7.385295e-02  6.285854e-06
   1.125088e-01 -1.194895e-05  1.371771e-06 -1.194895e-05  1.183999e-01 -1.194895e-05
  -2.123527e-02 -2.537410e-07 -2.129906e-07 -2.537410e-07 -1.910550e-02 -2.537410e-07
   3.640939e-02  3.214758e-07  2.178567e-06  3.214758e-07 -5.139584e-02  3.214758e-07
   9.305814e-02 -5.376336e-06  1.158183e-07 -5.376336e-06 -9.337402e-04 -5.376336e-06
   2.934651e-02 -4.049287e-06 -8.172477e-07 -4.049287e-06  7.213868e-02 -4.049287e-06
          [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -1.201027e-02  3.575219e-04  5.991798e-03  3.575219e-04  1.459972e-03  3.575219e-04
  -1.616030e-02  1.893807e-03 -5.413682e-02  1.893807e-03  6.139140e-02  1.893807e-03
  -6.150504e-07 -1.024752e-06 -1.265386e-05 -1.024752e-06 -2.071258e-06 -1.024752e-06
  -4.858236e-03  2.030337e-03  5.637752e-02  2.030337e-03 -1.469745e-01  2.030337e-03
  -5.775960e-02  2.358644e-04  5.221101e-02  2.358644e-04 -7.276791e-02  2.358644e-04
  -1.072021e-01 -3.345519e-03  2.185744e-02 -3.345519e-03 -9.028374e-02 -3.345519e-03
  -4.224940e-02 -4.827818e-04 -3.737540e-03 -4.827818e-04  1.759567e-02 -4.827818e-04
  -4.949194e-02 -4.384994e-04  4.147029e-02 -4.384994e-04  1.852035e-02 -4.384994e-04
  -4.730307e-02  2.714356e-04  2.192485e-02  2.714356e-04 -2.855575e-02  2.714356e-04
  -5.468679e-03  6.848355e-03  1.453524e-03  6.848355e-03  1.193266e-01  6.848355e-03
  -1.771742e-02  5.252320e-03  9.797827e-02  5.252320e-03  4.410153e-02  5.252320e-03
   1.247973e-02 -1.151208e-03  1.686217e-01 -1.151208e-03 -2.258599e-02 -1.151208e-03
   1.291672e-01  6.876345e-02  1.242513e+00  6.876345e-02  1.028768e-01  6.876345e-02
   3.707421e-02 -7.060139e-03  1.807025e-01 -7.060139e-03 -4.069938e-01 -7.060139e-03
  -3.584898e-03  7.647388e-03 -2.576808e-02  7.647388e-03 -1.274355e-03  7.647388e-03
   1.358725e-01 -9.622110e-03 -9.981120e-02 -9.622110e-03 -2.260521e-01 -9.622110e-03
  -9.396728e-03  1.458265e-02  2.164233e-01  1.458265e-02  1.858056e-01  1.458265e-02
   1.211209e-01  8.944856e-02  1.586626e+00  8.944856e-02  1.447643e-01  8.944856e-02
   4.141820e-02 -2.298850e-02 -2.146522e-01 -2.298850e-02  3.737647e-01 -2.298850e-02
  -2.826044e-02 -2.842588e-06  1.873720e-02 -2.842588e-06  1.225322e-01 -2.842588e-06
   2.836610e-02  2.668043e-02  8.316054e-01  2.668043e-02 -3.552386e-02  2.668043e-02
   3.900724e-02 -1.605477e-02 -6.881064e-02 -1.605477e-02  2.453334e-01 -1.605477e-02
  -1.690925e-01 -3.207616e-02 -6.146330e-01 -3.207616e-02  2.494164e-01 -3.207616e-02
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]
   1.577319e-02 -2.751835e-04 -1.000108e-03 -2.751835e-04 -3.434716e-03 -2.751835e-04
  -1.259132e-02 -5.690304e-03 -2.238120e-02 -5.690304e-03 -1.188775e-02 -5.690304e-03
  -2.373815e-06  1.029718e-07 -1.671917e-05  1.029718e-07 -2.918244e-06  1.029718e-07
   8.362720e-02  2.313893e-03  6.035128e-02  2.313893e-03  2.122411e-02  2.313893e-03
   9.492845e-02  7.694474e-03  1.255966e-01  7.694474e-03  4.262244e-03  7.694474e-03
   1.187391e-01  1.153349e-02 -5.436739e-02  1.153349e-02  1.669989e-02  1.153349e-02
  -3.412461e-03  1.576106e-03  7.670558e-03  1.576106e-03  9.425288e-03  1.576106e-03
   6.458537e-03  4.180304e-05  2.908242e-02  4.180304e-05  5.071690e-02  4.180304e-05
   8.876498e-02 -2.081839e-03  4.936132e-02 -2.081839e-03 -3.091464e-02 -2.081839e-03
   8.792320e-02  4.636077e-03  9.053258e-02  4.636077e-03  2.238453e-02  4.636077e-03
  -6.303788e-02  7.740886e-03  4.573763e-02  7.740886e-03  4.939773e-02  7.740886e-03
  -1.993041e-02  4.651129e-03  9.667978e-03  4.651129e-03  4.201227e-02  4.651129e-03
   1.409098e-01  4.241670e-02  1.393120e+00  4.241670e-02  2.900452e-01  4.241670e-02
   2.332679e-01 -2.153071e-02  5.126337e-01 -2.153071e-02  6.951312e-02 -2.153071e-02
  -1.142538e-02  1.509765e-02  8.690495e-02  1.509765e-02 -7.460638e-02  1.509765e-02
   6.108820e-01 -1.268221e-02  3.518212e-01 -1.268221e-02  8.051495e-02 -1.268221e-02
   2.640660e-01 -2.724879e-03  2.692114e-01 -2.724879e-03  8.715751e-02 -2.724879e-03
   1.244480e-01  7.084991e-02  9.191380e-01  7.084992e-02  3.940943e-01  7.084992e-02
   2.462195e+00  1.194610e-01 -2.940413e-01  1.194610e-01  7.824849e-01  1.194610e-01
   1.152969e-01  3.346237e-02  1.493990e-01  3.346237e-02 -1.479669e-01  3.346237e-02
   1.052722e-01  4.979546e-02  7.842352e-01  4.979546e-02 -2.052034e-03  4.979546e-02
   4.932294e-01 -7.583424e-03 -1.134310e-01 -7.583424e-03  1.580448e-01 -7.583424e-03
   1.012813e-01  5.779883e-02  7.477003e-02  5.779883e-02  4.934622e-01  5.779883e-02
          [,43]
  -3.531617e+00
  -1.462751e+00
  -3.644364e-05
  -5.332093e+00
  -3.567155e+00
  -7.717775e+00
  -1.801526e+00
  -3.636457e+00
  -3.366687e+00
  -4.488497e+00
  -1.522955e+00
  -1.737318e+00
  -5.553566e+01
  -1.945388e+01
  -2.557287e+00
  -3.923389e+01
  -1.180842e+01
  -3.540107e+01
  -5.825920e+01
  -6.572260e+00
  -1.102390e+01
  -2.883684e+01
  -2.579244e+01
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  8663 / 8663 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value  
s(measurement.no)                            9.00e+00 8.06e+00    0.99   0.225  
s(f0Zscore2)                                 9.00e+00 3.93e+00    1.04   1.000  
s(durationZscore2)                           2.90e+01 1.00e+00    0.98   0.080 .
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 3.42e+01    1.00   0.485  
ti(measurement.no,durationZscore2)           4.40e+01 1.19e+01    1.03   0.990  
ti(measurement.no,f0Zscore2)                 2.80e+01 1.67e+01    1.01   0.660  
ti(f0Zscore2,durationZscore2)                7.70e+01 7.52e+00    0.99   0.120  
s(word)                                      2.03e+02 1.11e+02      NA      NA  
s(wordPos)                                   3.09e+02 3.89e+01      NA      NA  
s(wordLeftRightTone)                         2.90e+02 5.11e+00      NA      NA  
s(measurement.no,wordPos)                    3.09e+02 7.85e+01      NA      NA  
s(f0Zscore2,wordPos)                         3.09e+02 2.36e+01      NA      NA  
s(durationZscore2,wordPos)                   3.09e+02 7.08e+01      NA      NA  
s(measurement.no,wordLeftRightTone)          2.90e+02 1.17e+02      NA      NA  
s(f0Zscore2,wordLeftRightTone)               2.90e+02 1.31e+01      NA      NA  
s(durationZscore2,wordLeftRightTone)         2.90e+02 2.20e+01      NA      NA  
s(measurement.no,word)                       2.03e+02 5.77e+01      NA      NA  
s(f0Zscore2,word)                            2.03e+02 5.16e+01      NA      NA  
s(durationZscore2,word)                      2.03e+02 3.98e-03      NA      NA  
s(measurement.no,speaker)                    1.00e+02 5.77e+01    0.99   0.190  
s(durationZscore2,speaker)                   3.00e+02 8.78e-04    0.98   0.095 .
s(f0Zscore2,speaker)                         1.00e+02 3.48e+01    1.04   0.995  
s(measurement.no,speakerPos)                 3.00e+02 7.22e+01    0.99   0.190  
s(durationZscore2,speakerPos)                9.00e+02 1.67e+02    0.98   0.080 .
s(f0Zscore2,speakerPos)                      3.00e+02 6.79e+01    1.04   1.000  
s(measurement.no,speakerLeftRightTone)       5.90e+02 1.13e+02    0.99   0.185  
s(durationZscore2,speakerLeftRightTone)      1.77e+03 1.66e+02    0.98   0.095 .
s(f0Zscore2,speakerLeftRightTone)            5.90e+02 9.40e+01    1.04   1.000  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

duration0.25f = quantile(data.ai.fem$durationZscore2, c(0.25)) * global_sddf + global_meandf

duration0.5f = quantile(data.ai.fem$durationZscore2, c(0.5)) * global_sddf + global_meandf

duration0.75f = quantile(data.ai.fem$durationZscore2, c(0.75)) * global_sddf + global_meandf

# 3D plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.627777955022104. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.ai.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0952338520379409. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.525044865601105. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2b, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.ai.fem$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.ai.fem$durationZscore2, c(0.05,0.95), na.rm = T),
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddf + global_meandf), 
           #zlim = c(-0.5,0.5),
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 30 values ranging from -1.397262 to 1.834229. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")

# 3D plotting
 png("pred2-4.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.627777955022104. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-5.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.ai.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0952338520379409. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-6.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2b, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1f + global_mean1f, 
        zlim = c(500,1000),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.525044865601105. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,1000), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-diff-2.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2b, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.ai.fem$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.ai.fem$durationZscore2, c(0.05,0.95), na.rm = T),
           n.grid = 100,
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddf + global_meandf), 
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 100 values ranging from -1.397262 to 1.834229. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")
dev.off()

null device 
          1 

5.5 Model 2C: /au/ male, F1 as output

system.time(gamm.model2c.noAR <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2c.noAR, paste("Gamm_model2c_noAR.rds"))

gamm.model2c.noAR <- 
  readRDS("Gamm_model2c_noAR.rds")

r.gamm.model2c <- start_value_rho(gamm.model2c.noAR)

# Auto-regressive model

system.time(gamm.model2c <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
 
                        
                    
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2c, AR.start = data.au.mas$start)
)

#saveRDS(gamm.model2c, paste("Gamm_model2c.rds"))

gamm.model2c <- 
  readRDS("Gamm_model2c.rds")

summary(gamm.model2c, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)  
(Intercept) -0.11603    0.05027  -2.308    0.021 *
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Approximate significance of smooth terms:
                                                edf  Ref.df      F p-value    
s(measurement.no)                             7.877   8.200 75.621  <2e-16 ***
s(f0Zscore2)                                  1.000   1.000  2.632   0.105    
s(durationZscore2)                            1.000   1.000  0.081   0.776    
ti(measurement.no,f0Zscore2,durationZscore2) 77.986 103.329  1.889  <2e-16 ***
ti(measurement.no,durationZscore2)           18.835  21.871  3.609  <2e-16 ***
ti(measurement.no,f0Zscore2)                  1.001   1.002  0.710   0.400    
ti(f0Zscore2,durationZscore2)                12.782  16.359  1.115   0.294    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.704   Deviance explained = 75.3%
fREML =   9240  Scale est. = 0.27859   n = 9522

gam.check(gamm.model2c)

Method: fREML   Optimizer: perf chol
$grad
 [1]  1.509903e-14 -4.421812e-05 -5.173769e-05 -5.329071e-15  2.913225e-13  5.719869e-13
 [7] -6.661338e-14 -1.625367e-13 -4.137722e-05 -7.430545e-05 -2.797762e-14  8.881784e-16
[13] -3.019807e-13  4.263256e-14  8.810730e-13 -1.989520e-13  1.065814e-13 -2.806644e-13
[19]  1.421085e-14 -7.744916e-13  3.694822e-13  6.075140e-13 -5.943035e-05  8.615331e-14
[25] -3.552714e-13 -4.513057e-14  1.332268e-13  2.009504e-14  9.592327e-14 -5.551115e-15
[31] -7.105427e-15 -2.216005e-13 -6.863694e-05  1.048051e-13 -2.096101e-13  3.108624e-14
[37] -7.105427e-14  8.659740e-15 -1.563194e-13 -2.642331e-14 -5.826450e-13 -4.507505e-14
[43]  6.366463e-12

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]
   3.436732e+00  1.146967e-07  2.882443e-07  1.981049e-02  5.047585e-03 -8.700809e-03
   1.146967e-07  4.421468e-05 -1.659474e-11 -5.374701e-07  1.115960e-06 -7.118915e-07
   2.882443e-07 -1.659474e-11  5.174411e-05  6.824430e-06 -4.923589e-06 -6.712126e-06
   1.981049e-02 -5.374701e-07  6.824430e-06  3.417803e+00  3.066756e-02  5.502142e-01
   5.047585e-03  1.115960e-06 -4.923589e-06  3.066756e-02  2.327217e+00  7.406141e-01
  -8.700809e-03 -7.118915e-07 -6.712126e-06  5.502142e-01  7.406141e-01  3.237080e+00
  -1.286566e-03  2.507969e-08  4.498085e-07 -4.576287e-03  1.492604e-03 -8.137149e-04
   8.758973e-03  6.527397e-08 -4.706856e-06  1.217636e-02  7.569456e-01  3.032156e-01
  -5.862305e-06 -2.051877e-10 -2.690992e-11  2.381049e-05 -2.370795e-06  5.321780e-06
  -2.365183e-06 -5.425586e-10  2.566724e-10  6.576532e-06 -3.111268e-06 -9.304259e-06
  -2.036204e-03 -1.843609e-07 -5.810039e-06  8.184781e-03 -1.453347e-01  5.335916e-02
  -1.605283e-03  7.029070e-08 -6.856684e-06  1.235986e-02 -2.783766e-02  2.790303e-02
  -5.044123e-03  8.504241e-07  3.420854e-06 -7.902620e-03  4.143748e-03  6.607474e-02
  -5.170657e-03  1.260296e-07  1.170420e-06 -1.010465e-02  7.225012e-02  2.643432e-02
  -3.765831e-03  7.069552e-07  7.810375e-07 -5.106029e-02  4.161716e-02  1.450322e-01
   9.824810e-03 -1.140226e-06  1.620483e-06 -7.117713e-03  3.114254e-02  1.858161e-01
   7.636842e-04 -7.422579e-07 -1.097526e-06 -1.301220e-02 -8.448967e-03 -3.785656e-02
   1.792185e-03 -3.364212e-07  3.255858e-07 -1.642050e-03 -3.348531e-02  9.556300e-03
   3.281171e-02 -2.462839e-06  2.262868e-05  1.455296e-01  2.173006e-01  2.586209e-01
   5.950808e-03 -4.126734e-07  1.206912e-06  1.286570e-01  1.635486e-01  1.661803e-01
  -2.423714e-03 -6.516604e-07  1.517338e-05 -4.596442e-03 -2.228014e-02 -5.952660e-03
   1.279447e-02 -8.036689e-07  1.176031e-05  1.845857e-02  7.865775e-03  1.095766e-01
   2.154697e-07 -4.678510e-11  3.379692e-10  5.635054e-07 -9.493168e-07  9.339618e-07
           [,7]          [,8]          [,9]         [,10]         [,11]         [,12]
  -1.286566e-03  8.758973e-03 -5.862305e-06 -2.365183e-06 -2.036204e-03 -1.605283e-03
   2.507969e-08  6.527397e-08 -2.051877e-10 -5.425586e-10 -1.843609e-07  7.029070e-08
   4.498085e-07 -4.706856e-06 -2.690992e-11  2.566724e-10 -5.810039e-06 -6.856684e-06
  -4.576287e-03  1.217636e-02  2.381049e-05  6.576532e-06  8.184781e-03  1.235986e-02
   1.492604e-03  7.569456e-01 -2.370795e-06 -3.111268e-06 -1.453347e-01 -2.783766e-02
  -8.137149e-04  3.032156e-01  5.321780e-06 -9.304259e-06  5.335916e-02  2.790303e-02
   1.235990e+00  2.098012e-02 -1.463403e-06  2.370865e-07  3.383424e-03  1.318383e-03
   2.098012e-02  2.460554e+00  4.547433e-06  4.118789e-06  3.683270e-02 -1.790341e-02
  -1.463403e-06  4.547433e-06  1.825315e-05  2.313124e-05 -2.393962e-06 -1.923176e-06
   2.370865e-07  4.118789e-06  2.313124e-05  5.115512e-05 -9.165367e-08 -1.135160e-06
   3.383424e-03  3.683270e-02 -2.393962e-06 -9.165367e-08  3.336908e-01  2.742291e-02
   1.318383e-03 -1.790341e-02 -1.923176e-06 -1.135160e-06  2.742291e-02  5.975948e-01
  -7.478653e-04  7.786644e-02 -8.355205e-07  3.282546e-06  1.844626e-02  4.152910e-02
  -1.096399e-03 -1.345364e-04 -6.305630e-06 -8.745094e-06  4.107130e-02  4.315394e-02
  -3.421797e-03  6.458722e-02  4.447359e-06  4.913114e-06  8.746249e-02  9.255732e-02
  -1.104824e-02  3.523019e-02  1.062062e-05  2.399194e-06  6.931733e-02  4.197034e-02
   1.262883e-04 -7.389672e-02  8.608780e-07 -1.002320e-05  1.788144e-02  2.724728e-02
   1.935456e-03 -1.325218e-02 -3.909039e-06 -2.721852e-06 -1.558030e-03  9.215226e-03
  -1.338227e-02  2.317473e-01  1.188044e-05 -1.712841e-05  9.812755e-02 -6.759184e-02
   1.621363e-03 -6.516093e-02  7.180822e-06 -5.675088e-06  8.736228e-02  5.570510e-02
   3.110299e-03 -4.185397e-02 -3.770343e-06 -3.311862e-06 -3.641578e-02  5.855254e-02
  -2.844596e-03  1.319172e-01 -2.949513e-06 -1.197719e-05  1.736657e-02 -3.308422e-02
   1.044315e-08 -1.008587e-06 -1.701829e-10 -3.338137e-10 -7.920326e-08 -7.798039e-07
          [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -5.044123e-03 -5.170657e-03 -3.765831e-03  9.824810e-03  7.636842e-04  1.792185e-03
   8.504241e-07  1.260296e-07  7.069552e-07 -1.140226e-06 -7.422579e-07 -3.364212e-07
   3.420854e-06  1.170420e-06  7.810375e-07  1.620483e-06 -1.097526e-06  3.255858e-07
  -7.902620e-03 -1.010465e-02 -5.106029e-02 -7.117713e-03 -1.301220e-02 -1.642050e-03
   4.143748e-03  7.225012e-02  4.161716e-02  3.114254e-02 -8.448967e-03 -3.348531e-02
   6.607474e-02  2.643432e-02  1.450322e-01  1.858161e-01 -3.785656e-02  9.556300e-03
  -7.478653e-04 -1.096399e-03 -3.421797e-03 -1.104824e-02  1.262883e-04  1.935456e-03
   7.786644e-02 -1.345364e-04  6.458722e-02  3.523019e-02 -7.389672e-02 -1.325218e-02
  -8.355205e-07 -6.305630e-06  4.447359e-06  1.062062e-05  8.608780e-07 -3.909039e-06
   3.282546e-06 -8.745094e-06  4.913114e-06  2.399194e-06 -1.002320e-05 -2.721852e-06
   1.844626e-02  4.107130e-02  8.746249e-02  6.931733e-02  1.788144e-02 -1.558030e-03
   4.152910e-02  4.315394e-02  9.255732e-02  4.197034e-02  2.724728e-02  9.215226e-03
   4.418184e+00  1.714447e+00  6.726898e+00 -3.780560e-01  6.024339e-01  1.902889e-01
   1.714447e+00  4.933483e+00  4.115198e+00 -9.747848e-01  3.984492e-01  3.839881e-01
   6.726898e+00  4.115198e+00  2.496460e+01 -8.019937e-01  1.290693e+00  1.013514e+00
  -3.780560e-01 -9.747848e-01 -8.019937e-01  1.404195e+01  1.434227e+00 -1.295892e-01
   6.024339e-01  3.984492e-01  1.290693e+00  1.434227e+00  6.299845e+00 -4.218351e-02
   1.902889e-01  3.839881e-01  1.013514e+00 -1.295892e-01 -4.218351e-02  1.451908e+00
  -1.525436e+00 -5.812829e-01 -4.405254e+00  6.073460e+00  3.009789e-01  7.592186e-02
   8.964667e-01  6.216553e-01  1.536906e+00  2.277931e-01  3.308957e+00  2.937632e-01
   7.870165e-01  7.543510e-01  3.905927e+00 -3.287134e-01  3.102669e-01  1.429415e+00
  -8.514886e-01 -1.948560e-01 -1.309466e+00  3.268214e+00  2.639827e-04 -3.958702e-03
   1.806656e-05  1.523651e-05  4.266632e-05  4.049010e-08  7.413654e-05  2.330967e-06
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]
   3.281171e-02  5.950808e-03 -2.423714e-03  1.279447e-02  2.154697e-07 -3.572752e-04
  -2.462839e-06 -4.126734e-07 -6.516604e-07 -8.036689e-07 -4.678510e-11 -7.339471e-08
   2.262868e-05  1.206912e-06  1.517338e-05  1.176031e-05  3.379692e-10  4.082134e-06
   1.455296e-01  1.286570e-01 -4.596442e-03  1.845857e-02  5.635054e-07  2.959160e-03
   2.173006e-01  1.635486e-01 -2.228014e-02  7.865775e-03 -9.493168e-07 -8.113518e-04
   2.586209e-01  1.661803e-01 -5.952660e-03  1.095766e-01  9.339618e-07  4.522631e-03
  -1.338227e-02  1.621363e-03  3.110299e-03 -2.844596e-03  1.044315e-08  4.272748e-04
   2.317473e-01 -6.516093e-02 -4.185397e-02  1.319172e-01 -1.008587e-06 -5.316706e-03
   1.188044e-05  7.180822e-06 -3.770343e-06 -2.949513e-06 -1.701829e-10 -3.939274e-07
  -1.712841e-05 -5.675088e-06 -3.311862e-06 -1.197719e-05 -3.338137e-10 -3.854477e-08
   9.812755e-02  8.736228e-02 -3.641578e-02  1.736657e-02 -7.920326e-08 -7.264632e-03
  -6.759184e-02  5.570510e-02  5.855254e-02 -3.308422e-02 -7.798039e-07  4.699685e-03
  -1.525436e+00  8.964667e-01  7.870165e-01 -8.514886e-01  1.806656e-05  1.390401e-01
  -5.812829e-01  6.216553e-01  7.543510e-01 -1.948560e-01  1.523651e-05  4.432182e-02
  -4.405254e+00  1.536906e+00  3.905927e+00 -1.309466e+00  4.266632e-05  2.289523e-01
   6.073460e+00  2.277931e-01 -3.287134e-01  3.268214e+00  4.049010e-08 -3.860953e-02
   3.009789e-01  3.308957e+00  3.102669e-01  2.639827e-04  7.413654e-05  8.287341e-03
   7.592186e-02  2.937632e-01  1.429415e+00 -3.958702e-03  2.330967e-06  2.951279e-01
   3.914752e+01  1.246009e+00 -3.577944e-01  8.532207e+00  1.574717e-05  1.282669e-02
   1.246009e+00  1.446358e+01  2.407992e-01  2.689129e-02  1.731932e-04  1.346013e-02
  -3.577944e-01  2.407992e-01  5.954430e+00 -2.572411e-02  7.813866e-06  5.232229e-01
   8.532207e+00  2.689129e-02 -2.572411e-02  5.124133e+00  2.355977e-06 -1.289592e-02
   1.574717e-05  1.731932e-04  7.813866e-06  2.355977e-06  5.943642e-05  7.173093e-07
          [,25]         [,26]         [,27]         [,28]         [,29]         [,30]
  -1.675342e-02 -3.426797e-04 -3.474186e-05 -3.426797e-04  5.900069e-04 -3.426797e-04
  -4.039322e-08  1.121224e-08 -5.304256e-07  1.121224e-08 -1.625384e-05  1.121224e-08
   2.888079e-06  9.577434e-07 -2.451360e-05  9.577434e-07 -2.105010e-06  9.577434e-07
   3.189405e-02  4.985454e-03 -8.172980e-03  4.985454e-03 -6.869081e-03  4.985454e-03
   3.985790e-02  7.431636e-03 -4.880679e-02  7.431636e-03  1.183344e-01  7.431636e-03
   1.338516e-01  4.389577e-03 -5.472110e-03  4.389577e-03 -8.360569e-02  4.389577e-03
  -1.193198e-02 -1.128676e-04  1.540164e-03 -1.128676e-04 -2.723718e-04 -1.128676e-04
   1.217512e-02 -4.750286e-04  8.015042e-03 -4.750286e-04  4.016303e-02 -4.750286e-04
  -1.131271e-06  4.737034e-08  1.142241e-06  4.737034e-08  1.258828e-05  4.737034e-08
   2.095554e-06  2.061933e-07  2.998105e-06  2.061933e-07  1.076834e-05  2.061933e-07
   1.287412e-02 -2.296304e-04  3.517107e-02 -2.296304e-04  2.155815e-02 -2.296304e-04
   9.819803e-03  2.997151e-04  1.436125e-02  2.997151e-04  1.324112e-01  2.997151e-04
   3.107447e-02  7.258587e-03  1.080013e-01  7.258587e-03  1.301775e-01  7.258587e-03
  -3.683349e-02 -3.031052e-03 -5.463816e-02 -3.031052e-03  1.771407e-01 -3.031052e-03
   1.464152e-01  2.567387e-02  2.008312e-02  2.567387e-02  2.255219e-01  2.567387e-02
  -8.838208e-02  2.057097e-02 -5.233080e-03  2.057097e-02  1.231381e-01  2.057097e-02
   6.025999e-02 -6.490604e-03  2.030272e-02 -6.490604e-03  2.696703e-01 -6.490604e-03
   2.207357e-02 -6.408460e-03  2.277430e-01 -6.408460e-03  7.451391e-03 -6.408460e-03
  -7.573968e-02 -2.120945e-02  2.651579e-01 -2.120945e-02  1.618256e-01 -2.120945e-02
   6.045905e-02  1.414884e-03  1.078129e-01  1.414884e-03  4.440862e-02  1.414884e-03
   6.948562e-02  5.913931e-03  5.143314e-01  5.913931e-03  4.410556e-02  5.913931e-03
  -3.788700e-02 -1.555894e-03  2.496553e-02 -1.555894e-03 -9.600259e-03 -1.555894e-03
   1.083062e-06 -1.844158e-07  2.431223e-06 -1.844158e-07 -2.052361e-06 -1.844158e-07
          [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -4.118181e-03  4.697898e-04  2.978866e-06  4.697898e-04 -6.569226e-03  4.697898e-04
   1.083401e-06  1.440190e-07 -9.052728e-11  1.440190e-07 -1.797516e-06  1.440190e-07
   3.744783e-07  2.728049e-06  2.156991e-09  2.728049e-06 -5.237113e-06  2.728049e-06
   3.389692e-02  1.383244e-02 -1.370360e-05  1.383244e-02 -3.856958e-02  1.383244e-02
   9.408878e-02  1.025509e-02 -8.450354e-06  1.025509e-02  9.472155e-02  1.025509e-02
  -1.214277e-01 -4.757924e-03  5.844013e-05 -4.757924e-03  5.123134e-02 -4.757924e-03
   1.463968e-05 -2.951403e-04  1.246718e-06 -2.951403e-04  2.776112e-03 -2.951403e-04
   1.280036e-01  1.291575e-03 -3.163894e-05  1.291575e-03  5.771936e-02  1.291575e-03
  -2.428938e-05 -5.142980e-07 -2.069469e-09 -5.142980e-07  2.828756e-07 -5.142980e-07
  -1.759840e-05  5.279717e-07  5.008835e-10  5.279717e-07 -3.800829e-06  5.279717e-07
  -7.138344e-02  8.089754e-03  1.118953e-05  8.089754e-03 -3.894895e-03  8.089754e-03
  -5.158876e-02  7.121909e-03  3.288369e-05  7.121909e-03  4.184381e-02  7.121909e-03
   3.960042e-02 -1.028662e-03  1.624656e-04 -1.028662e-03  2.386837e-01 -1.028662e-03
  -5.075032e-03 -2.932551e-02 -1.054670e-05 -2.932551e-02  1.821882e-01 -2.932551e-02
   6.637041e-02  9.118218e-02  3.088108e-04  9.118219e-02  3.457531e-01  9.118219e-02
  -2.468397e-01  3.652727e-02 -1.763624e-05  3.652727e-02  4.562811e-02  3.652727e-02
  -1.285211e-02  7.768605e-02  1.260467e-04  7.768605e-02  4.901127e-01  7.768605e-02
  -1.529099e-02 -6.430168e-03  4.048850e-06 -6.430168e-03  1.558641e-04 -6.430168e-03
  -4.421387e-01 -1.276759e-01  1.534479e-05 -1.276759e-01  5.359392e-02 -1.276759e-01
  -1.116422e-01  3.203428e-02  1.449917e-04  3.203428e-02  3.055103e-01  3.203428e-02
   2.974036e-02  1.406145e-02  4.797451e-04  1.406145e-02  1.584431e-01  1.406145e-02
   5.911798e-03 -6.617797e-03 -1.438953e-05 -6.617797e-03 -1.665766e-02 -6.617797e-03
   1.316931e-06 -8.564653e-07  2.282220e-09 -8.564654e-07 -6.947652e-07 -8.564654e-07
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]
  -2.167503e-02 -7.975354e-04 -3.949385e-03 -7.975354e-04 -8.273465e-03 -7.975354e-04
   2.771373e-07 -7.530748e-08 -6.445237e-07 -7.530748e-08 -2.961145e-06 -7.530748e-08
   3.355455e-06  1.948591e-06  4.821005e-06  1.948591e-06 -2.509218e-06  1.948591e-06
   3.519740e-01  6.781569e-03  1.609680e-02  6.781569e-03 -1.967145e-01  6.781569e-03
   2.445473e-01  9.472491e-03  4.031329e-02  9.472491e-03  1.658772e-01  9.472491e-03
   6.049784e-02  1.233805e-02 -5.986449e-02  1.233805e-02  1.744240e-01  1.233805e-02
  -5.480168e-02  2.550335e-04  1.201860e-02  2.550335e-04 -3.866488e-03  2.550335e-04
  -8.656911e-02 -3.387216e-03  3.734572e-02 -3.387216e-03 -5.526570e-02 -3.387216e-03
  -5.856668e-05 -1.047904e-06  7.314227e-06 -1.047904e-06 -4.937067e-07 -1.047904e-06
  -3.100401e-05 -1.335578e-07  1.016292e-05 -1.335578e-07 -9.720280e-06 -1.335578e-07
   1.441558e-01 -7.227866e-03  1.293271e-01 -7.227866e-03  1.742858e-02 -7.227866e-03
  -8.043607e-04  1.598544e-04  6.084896e-02  1.598544e-04  1.066139e-01  1.598544e-04
   3.978378e-01 -2.913428e-03  5.887078e-01 -2.913428e-03  1.156186e-01 -2.913428e-03
   1.431061e-01 -2.455773e-02  6.401911e-01 -2.455773e-02  2.420734e-01 -2.455773e-02
   3.113665e-01  7.232662e-02  2.357553e+00  7.232662e-02 -8.046154e-02  7.232662e-02
   1.115090e+00 -1.369217e-03 -3.422448e-01 -1.369217e-03  5.566745e-01 -1.369217e-03
   1.210017e-01 -3.405074e-02  6.129902e-01 -3.405074e-02  1.202833e+00 -3.405074e-02
  -3.209561e-02  1.081309e-02  1.554538e+00  1.081309e-02 -6.915617e-02  1.081309e-02
   6.991887e+00  3.765918e-02  2.211238e-01  3.765918e-02  8.668306e-01  3.765918e-02
  -9.291237e-02 -3.954164e-02  1.249711e+00 -3.954164e-02 -3.559875e-01 -3.954164e-02
   1.180762e-01  1.385454e-02  1.436396e+00  1.385454e-02 -2.597710e-01  1.385454e-02
   1.030678e+00 -6.382651e-03  5.136963e-03 -6.382651e-03  2.263691e-01 -6.382651e-03
  -5.650191e-06 -8.451229e-08  3.475033e-05 -8.451229e-08  1.879795e-05 -8.451229e-08
          [,43]
  -3.438352e+00
  -2.184837e-05
  -9.596625e-05
  -1.578435e+01
  -1.062259e+01
  -1.208621e+01
  -2.487165e+00
  -6.430418e+00
  -1.980775e-04
  -2.254982e-04
  -3.009853e+00
  -2.881024e+00
  -2.423717e+01
  -2.426736e+01
  -7.428494e+01
  -5.096893e+01
  -3.088692e+01
  -1.148205e+01
  -9.470367e+01
  -4.353874e+01
  -2.590247e+01
  -2.594514e+01
  -7.427970e-04
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  10417 / 10417 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value
s(measurement.no)                            9.00e+00 7.88e+00    1.03    0.97
s(f0Zscore2)                                 9.00e+00 1.00e+00    1.00    0.44
s(durationZscore2)                           2.90e+01 1.00e+00    1.01    0.69
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 7.80e+01    1.00    0.52
ti(measurement.no,durationZscore2)           4.40e+01 1.88e+01    1.02    0.89
ti(measurement.no,f0Zscore2)                 2.80e+01 1.00e+00    1.00    0.58
ti(f0Zscore2,durationZscore2)                7.70e+01 1.28e+01    1.02    0.82
s(word)                                      3.19e+02 4.85e+01      NA      NA
s(wordPos)                                   4.66e+02 4.85e+01      NA      NA
s(wordLeftRightTone)                         4.43e+02 1.49e+02      NA      NA
s(measurement.no,wordPos)                    4.66e+02 1.02e+02      NA      NA
s(f0Zscore2,wordPos)                         4.66e+02 6.18e+01      NA      NA
s(durationZscore2,wordPos)                   4.66e+02 2.30e+01      NA      NA
s(measurement.no,wordLeftRightTone)          4.43e+02 1.89e+02      NA      NA
s(f0Zscore2,wordLeftRightTone)               4.43e+02 8.71e+01      NA      NA
s(durationZscore2,wordLeftRightTone)         4.43e+02 5.18e+01      NA      NA
s(measurement.no,word)                       3.19e+02 5.19e+01      NA      NA
s(f0Zscore2,word)                            3.19e+02 1.60e-03      NA      NA
s(durationZscore2,word)                      3.19e+02 4.88e+00      NA      NA
s(measurement.no,speaker)                    1.00e+02 3.70e+01    1.03    0.98
s(durationZscore2,speaker)                   3.00e+02 3.19e+01    1.01    0.71
s(f0Zscore2,speaker)                         1.00e+02 3.83e+01    1.00    0.39
s(measurement.no,speakerPos)                 3.00e+02 6.18e+01    1.03    0.99
s(durationZscore2,speakerPos)                9.00e+02 4.10e+00    1.01    0.71
s(f0Zscore2,speakerPos)                      3.00e+02 4.09e+01    1.00    0.43
s(measurement.no,speakerLeftRightTone)       6.00e+02 1.55e+02    1.03    0.97
s(durationZscore2,speakerLeftRightTone)      1.80e+03 1.64e+02    1.01    0.68
s(f0Zscore2,speakerLeftRightTone)            6.00e+02 1.12e+02    1.00    0.43

duration0.25au = quantile(data.au.mas$durationZscore2, c(0.25)) * global_sddau + global_meandau

duration0.5au = quantile(data.au.mas$durationZscore2, c(0.5)) * global_sddau + global_meandau

duration0.75au = quantile(data.au.mas$durationZscore2, c(0.75)) * global_sddau + global_meandau

# 3D plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.680380067515762. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0990769355270091. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.530017553918074. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2c, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.au.mas$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.au.mas$durationZscore2, c(0.05,0.95), na.rm = T),
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddau + global_meandau), 
           #zlim = c(-0.5,0.5),
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 30 values ranging from -1.441948 to 1.867158. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")

# 3D plotting
 png("pred2-7.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.680380067515762. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2-8.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0990769355270091. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2-9.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2c, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1au + global_mean1au, 
        zlim = c(490,720),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.530017553918074. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(490,720), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-diff-3.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2c, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.au.mas$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.au.mas$durationZscore2, c(0.05,0.95), na.rm = T),
           n.grid = 100,
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddau + global_meandau), 
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 100 values ranging from -1.441948 to 1.867158. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")
dev.off()

null device 
          1 

5.6 Model 2D: /au/ female, F1 as output

system.time(gamm.model2d.noAR <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2d.noAR, paste("Gamm_model2d_noAR.rds"))

gamm.model2d.noAR <- 
  readRDS("Gamm_model2d_noAR.rds")

r.gamm.model2d <- start_value_rho(gamm.model2d.noAR)

# Auto-regressive model

system.time(gamm.model2d <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
 
                        
                    
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2d, AR.start = data.au.fem$start)
)

#saveRDS(gamm.model2d, paste("Gamm_model2d.rds"))

gamm.model2d <- 
  readRDS("Gamm_model2d.rds")

summary(gamm.model2d, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept)  0.01375    0.04473   0.307    0.759

Approximate significance of smooth terms:
                                                edf Ref.df       F p-value    
s(measurement.no)                             8.169  8.388 122.102  <2e-16 ***
s(f0Zscore2)                                  1.996  2.172   0.769   0.523    
s(durationZscore2)                            1.000  1.000   0.358   0.550    
ti(measurement.no,f0Zscore2,durationZscore2) 28.296 42.725   0.934   0.597    
ti(measurement.no,durationZscore2)           11.748 14.646   5.500  <2e-16 ***
ti(measurement.no,f0Zscore2)                  9.712 12.616   6.688  <2e-16 ***
ti(f0Zscore2,durationZscore2)                12.318 15.965   1.024   0.429    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.763   Deviance explained = 79.9%
fREML = 7714.8  Scale est. = 0.22484   n = 9332

gam.check(gamm.model2d)

Method: fREML   Optimizer: perf chol
$grad
 [1] -7.638334e-14 -3.141931e-14 -3.569072e-05  2.842171e-14 -9.858780e-14 -5.684342e-14
 [7] -1.798561e-14  1.585398e-13  1.549871e-13  1.794120e-13  8.393286e-14 -1.207923e-13
[13]  9.947598e-14 -5.263191e-05 -6.483702e-14  3.836931e-13 -1.243450e-13 -9.059420e-14
[19] -6.963319e-13 -2.842171e-14  9.237056e-14  5.258016e-13 -5.385358e-05 -6.177734e-05
[25]  1.598721e-13 -3.620603e-05 -1.723066e-13 -3.620603e-05  5.009326e-13 -3.620603e-05
[31] -3.268497e-13  7.194245e-14  2.273737e-13 -3.530509e-14 -1.207923e-13 -8.970602e-14
[37]  3.979039e-13  1.012523e-13  1.278977e-13  7.904788e-14  1.492140e-13 -2.753353e-13
[43] -3.183231e-11

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]
   3.693811e+00 -1.293109e-03  7.074668e-09  6.300282e-03 -2.113894e-03 -1.858520e-03
  -1.293109e-03  2.194398e-01 -1.300687e-07  1.434675e-03  5.037439e-03 -1.865275e-03
   7.074668e-09 -1.300687e-07  3.568931e-05 -2.014656e-07 -2.062342e-07 -2.340776e-07
   6.300282e-03  1.434675e-03 -2.014656e-07  1.398911e+00 -7.652754e-02 -1.340451e-02
  -2.113894e-03  5.037439e-03 -2.062342e-07 -7.652754e-02  7.858723e-01  1.452559e-01
  -1.858520e-03 -1.865275e-03 -2.340776e-07 -1.340451e-02  1.452559e-01  7.364177e-01
  -3.537795e-03 -1.622780e-04 -1.601955e-07  4.959838e-02 -2.421809e-03  5.780542e-04
   2.630167e-03 -5.084841e-03 -1.096702e-06 -4.927767e-02  1.414499e-01  3.834754e-02
   8.150346e-03 -4.105654e-02 -4.627279e-08 -1.865503e-02 -8.700575e-03 -1.126091e-02
   2.242578e-03  2.077654e-02  2.843806e-08 -1.493708e-02 -1.364056e-02 -5.340625e-03
  -2.531066e-04 -3.035689e-03 -1.760637e-06 -3.594216e-02 -7.968544e-03  4.059835e-02
  -1.404753e-03 -5.531223e-03 -1.672625e-06  3.456931e-03  1.446608e-02  1.064430e-01
  -2.825061e-02 -6.197482e-04 -1.486449e-06  4.183666e-02 -1.464845e-02  2.632387e-03
  -1.132987e-07 -4.145418e-07  1.165927e-11  1.434485e-06 -3.626290e-06 -3.993923e-06
  -2.287224e-03  4.470919e-04  2.448806e-09  7.466808e-03  3.405113e-03  5.052973e-03
   3.823419e-03  2.821802e-02 -3.143066e-07  5.604196e-02 -3.314030e-02 -6.214901e-02
   7.307307e-03  2.291743e-03 -1.349935e-07  2.754456e-02  4.878053e-02  2.903468e-02
  -1.769562e-04  2.680812e-03 -3.009771e-06 -1.257206e-02 -4.579390e-04  2.732203e-03
   7.049977e-03  1.281553e-01 -2.776881e-07  6.063706e-02  1.744228e-02  3.830774e-03
   1.175073e-02  1.438189e-03 -4.978680e-08 -4.996891e-02  1.263396e-02  8.229128e-02
   1.710928e-04 -4.627713e-03 -4.084055e-06  1.580290e-02  2.407712e-02  1.294446e-02
   8.471170e-03  3.480302e-02 -2.645565e-08  4.756441e-02  1.542074e-02  5.357954e-02
   2.032989e-06 -4.571933e-06 -9.543631e-11 -9.925387e-06  1.135209e-05  5.842666e-06
           [,7]          [,8]          [,9]         [,10]         [,11]         [,12]
  -3.537795e-03  2.630167e-03  8.150346e-03  2.242578e-03 -2.531066e-04 -1.404753e-03
  -1.622780e-04 -5.084841e-03 -4.105654e-02  2.077654e-02 -3.035689e-03 -5.531223e-03
  -1.601955e-07 -1.096702e-06 -4.627279e-08  2.843806e-08 -1.760637e-06 -1.672625e-06
   4.959838e-02 -4.927767e-02 -1.865503e-02 -1.493708e-02 -3.594216e-02  3.456931e-03
  -2.421809e-03  1.414499e-01 -8.700575e-03 -1.364056e-02 -7.968544e-03  1.446608e-02
   5.780542e-04  3.834754e-02 -1.126091e-02 -5.340625e-03  4.059835e-02  1.064430e-01
   8.029237e-01  5.159074e-01  7.038664e-03  8.976371e-03 -5.193109e-03 -2.022717e-02
   5.159074e-01  1.187402e+00 -2.911643e-02  4.740459e-03  1.036725e-02 -1.382336e-01
   7.038664e-03 -2.911643e-02  7.062269e-01  3.257803e-01 -1.319765e-03  1.904896e-03
   8.976371e-03  4.740459e-03  3.257803e-01  1.830347e+00  9.775003e-03  5.473488e-05
  -5.193109e-03  1.036725e-02 -1.319765e-03  9.775003e-03  2.017259e-01  9.277563e-02
  -2.022717e-02 -1.382336e-01  1.904896e-03  5.473488e-05  9.277563e-02  4.921835e-01
   7.815647e-04 -6.659160e-03  1.805454e-02 -4.233450e-02  5.284732e-02  6.780803e-02
   2.689860e-07 -9.479245e-08  6.508174e-07  2.638659e-06  3.945202e-06  4.142837e-06
   1.732338e-03  1.699026e-03  2.415114e-03 -2.847150e-03  7.683969e-03  4.134518e-04
   1.481188e-02  8.436779e-02  3.940296e-02  2.181303e-01 -9.222560e-02  1.277405e-02
   1.880453e-03  2.286371e-02 -3.329043e-02  1.529037e-02 -1.553589e-02  1.794541e-02
   7.961150e-03  3.338596e-02 -1.228551e-03 -1.933922e-03  1.747518e-02  3.251301e-02
   1.477099e-03  2.945211e-02  1.967571e-02  4.780138e-02  1.004379e-02 -1.222640e-02
   3.538243e-02 -2.409386e-02  3.762773e-02  7.397851e-02  2.100707e-02  1.666081e-01
   7.608580e-03  8.868046e-03 -4.963780e-03  1.688211e-02  7.185994e-03  2.988403e-02
  -1.197748e-02  2.465842e-02  4.574251e-02 -1.661694e-02  1.798724e-02 -5.961522e-02
   1.088793e-05 -9.811830e-07  1.222769e-05 -2.892372e-07  4.262671e-06  2.349473e-05
          [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -2.825061e-02 -1.132987e-07 -2.287224e-03  3.823419e-03  7.307307e-03 -1.769562e-04
  -6.197482e-04 -4.145418e-07  4.470919e-04  2.821802e-02  2.291743e-03  2.680812e-03
  -1.486449e-06  1.165927e-11  2.448806e-09 -3.143066e-07 -1.349935e-07 -3.009771e-06
   4.183666e-02  1.434485e-06  7.466808e-03  5.604196e-02  2.754456e-02 -1.257206e-02
  -1.464845e-02 -3.626290e-06  3.405113e-03 -3.314030e-02  4.878053e-02 -4.579390e-04
   2.632387e-03 -3.993923e-06  5.052973e-03 -6.214901e-02  2.903468e-02  2.732203e-03
   7.815647e-04  2.689860e-07  1.732338e-03  1.481188e-02  1.880453e-03  7.961150e-03
  -6.659160e-03 -9.479245e-08  1.699026e-03  8.436779e-02  2.286371e-02  3.338596e-02
   1.805454e-02  6.508174e-07  2.415114e-03  3.940296e-02 -3.329043e-02 -1.228551e-03
  -4.233450e-02  2.638659e-06 -2.847150e-03  2.181303e-01  1.529037e-02 -1.933922e-03
   5.284732e-02  3.945202e-06  7.683969e-03 -9.222560e-02 -1.553589e-02  1.747518e-02
   6.780803e-02  4.142837e-06  4.134518e-04  1.277405e-02  1.794541e-02  3.251301e-02
   4.121370e+01  2.098707e-04  1.346828e+00 -1.970201e+00  8.467952e-01  1.796074e+00
   2.098707e-04  5.264541e-05  1.709625e-05  5.560039e-05  5.408541e-05  3.760439e-05
   1.346828e+00  1.709625e-05  1.981955e-01 -9.170871e-02  8.497324e-02  8.725251e-02
  -1.970201e+00  5.560039e-05 -9.170871e-02  2.484061e+01  1.068712e+00  2.437450e-01
   8.467952e-01  5.408541e-05  8.497324e-02  1.068712e+00  3.954661e+00  7.034213e-02
   1.796074e+00  3.760439e-05  8.725251e-02  2.437450e-01  7.034213e-02  2.775935e+00
  -2.790637e+00 -2.920820e-05  9.940070e-03  7.053991e+00  5.585878e-01 -1.736453e-01
   1.789642e+00 -2.009365e-05  1.242494e-01  1.206986e+00  3.852502e+00  1.301961e-01
   1.917280e+00  3.752566e-05  1.222239e-01  7.162070e-01  1.811268e-01  1.112223e+00
  -3.329156e+00 -2.759203e-05 -7.698211e-02  6.903767e+00  3.129644e-01 -1.064590e-01
   3.059248e-04 -7.639956e-09  9.085043e-06  1.787077e-04  6.572387e-04  4.128209e-06
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]
   7.049977e-03  1.175073e-02  1.710928e-04  8.471170e-03  2.032989e-06  6.605885e-08
   1.281553e-01  1.438189e-03 -4.627713e-03  3.480302e-02 -4.571933e-06  1.912264e-09
  -2.776881e-07 -4.978680e-08 -4.084055e-06 -2.645565e-08 -9.543631e-11 -1.498799e-10
   6.063706e-02 -4.996891e-02  1.580290e-02  4.756441e-02 -9.925387e-06  1.524397e-06
   1.744228e-02  1.263396e-02  2.407712e-02  1.542074e-02  1.135209e-05  2.001639e-06
   3.830774e-03  8.229128e-02  1.294446e-02  5.357954e-02  5.842666e-06  1.532361e-08
   1.477099e-03  3.538243e-02  7.608580e-03 -1.197748e-02  1.088793e-05  3.667870e-07
   2.945211e-02 -2.409386e-02  8.868046e-03  2.465842e-02 -9.811830e-07  1.483118e-06
   1.967571e-02  3.762773e-02 -4.963780e-03  4.574251e-02  1.222769e-05 -4.275927e-07
   4.780138e-02  7.397851e-02  1.688211e-02 -1.661694e-02 -2.892372e-07  3.242257e-07
   1.004379e-02  2.100707e-02  7.185994e-03  1.798724e-02  4.262671e-06  1.475583e-06
  -1.222640e-02  1.666081e-01  2.988403e-02 -5.961522e-02  2.349473e-05  1.631695e-06
  -2.790637e+00  1.789642e+00  1.917280e+00 -3.329156e+00  3.059248e-04  1.141830e-04
  -2.920820e-05 -2.009365e-05  3.752566e-05 -2.759203e-05 -7.639956e-09  1.414896e-09
   9.940070e-03  1.242494e-01  1.222239e-01 -7.698211e-02  9.085043e-06  4.603143e-06
   7.053991e+00  1.206986e+00  7.162070e-01  6.903767e+00  1.787077e-04  4.418792e-05
   5.585878e-01  3.852502e+00  1.811268e-01  3.129644e-01  6.572387e-04  5.469031e-06
  -1.736453e-01  1.301961e-01  1.112223e+00 -1.064590e-01  4.128209e-06  4.828541e-05
   2.432811e+01  1.385623e+00  4.537777e-02  9.535978e+00  2.159451e-04  5.801488e-06
   1.385623e+00  1.583574e+01  2.864884e-01  5.912890e-01  1.804665e-03  1.235761e-05
   4.537777e-02  2.864884e-01  3.265159e+00 -1.765892e-01  3.689874e-05  8.612842e-05
   9.535978e+00  5.912890e-01 -1.765892e-01  1.143712e+01  9.888860e-05 -6.456625e-06
   2.159451e-04  1.804665e-03  3.689874e-05  9.888860e-05  5.422293e-05  1.475226e-09
          [,25]         [,26]         [,27]         [,28]         [,29]         [,30]
  -4.203142e-02  1.018275e-06  2.072727e-03  1.018275e-06 -4.372825e-03  1.018275e-06
  -3.027832e-03  2.848966e-08 -8.939348e-03  2.848966e-08  8.293042e-02  2.848966e-08
  -5.668779e-08 -8.723724e-11 -5.848275e-06 -8.723724e-11 -8.134251e-08 -8.723724e-11
  -2.544738e-02 -4.127585e-06  4.282252e-03 -4.127585e-06 -2.811938e-02 -4.127585e-06
   5.467154e-03  2.673399e-06 -1.352331e-02  2.673399e-06 -6.134909e-03  2.673399e-06
  -2.427352e-02  9.063456e-06 -3.058686e-02  9.063456e-06 -5.621221e-02  9.063456e-06
  -2.602896e-02  3.862345e-06 -2.980062e-03  3.862345e-06 -1.426849e-02  3.862345e-06
  -1.773744e-02 -5.404395e-06  1.777407e-03 -5.404395e-06 -2.033902e-02 -5.404395e-06
  -2.193403e-02 -9.614784e-07 -8.427449e-03 -9.614784e-07 -5.290427e-02 -9.614784e-07
  -6.405357e-03 -2.996745e-06  9.485546e-03 -2.996745e-06  2.633004e-02 -2.996745e-06
   2.061423e-03  2.533949e-06 -9.488510e-03  2.533949e-06  5.006321e-02  2.533949e-06
  -1.805444e-02  5.437235e-06 -5.568753e-02  5.437235e-06  2.090253e-02  5.437235e-06
   3.317148e-02  5.117850e-05  2.359596e-01  5.117850e-05 -4.153107e-03  5.117850e-05
   2.707820e-06  1.893742e-10  6.664545e-06  1.893742e-10 -1.559287e-06  1.893742e-10
   1.737144e-03  1.468346e-06  3.180275e-02  1.468346e-06  9.978841e-03  1.468346e-06
   1.293779e-01 -5.722926e-06 -6.767317e-02 -5.722926e-06  4.166987e-01 -5.722926e-06
   7.676558e-03  5.178052e-05  1.151279e-01  5.178052e-05  1.895511e-01  5.178052e-05
   1.500117e-02  5.315529e-06  1.674212e-01  5.315529e-06 -1.450386e-01  5.315529e-06
   1.379741e-02 -1.261651e-04 -1.566236e-01 -1.261651e-04  1.548172e-01 -1.261651e-04
  -2.017468e-02  2.315930e-05  9.968562e-02  2.315930e-05  2.033363e-01  2.315930e-05
   2.495518e-02 -1.299168e-05  3.123129e-01 -1.299168e-05  9.820455e-02 -1.299168e-05
   3.604611e-02 -2.794716e-05 -1.768277e-02 -2.794716e-05  3.769855e-02 -2.794716e-05
  -1.094904e-05 -5.619199e-09 -1.249313e-05 -5.619199e-09 -5.550326e-05 -5.619199e-09
          [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -1.385982e-02  3.389515e-04  7.726551e-03  3.389515e-04  7.103197e-04  3.389515e-04
   1.402268e-02 -6.271047e-04 -3.498037e-02 -6.271047e-04  9.269839e-03 -6.271047e-04
   1.257083e-07 -6.065322e-09 -6.884602e-06 -6.065322e-09 -2.350812e-07 -6.065322e-09
   1.161764e-01 -6.511501e-04  1.723055e-03 -6.511501e-04 -3.587591e-02 -6.511501e-04
   5.964604e-02  8.710931e-04 -1.932493e-02  8.710931e-04  1.906095e-02  8.710931e-04
  -2.108326e-02  3.500532e-03 -1.143800e-02  3.500532e-03 -2.009814e-02  3.500532e-03
   2.899256e-03 -1.386158e-03  1.789674e-02 -1.386158e-03 -4.912638e-03 -1.386158e-03
   3.701481e-02 -1.166805e-03 -3.960796e-02 -1.166805e-03 -2.228652e-02 -1.166805e-03
   2.660998e-02  1.048228e-03 -2.140103e-02  1.048228e-03 -3.224757e-02  1.048228e-03
  -2.211772e-02 -2.359295e-03  5.337264e-02 -2.359295e-03 -3.879344e-02 -2.359295e-03
   5.322684e-02 -1.479645e-03 -1.028497e-02 -1.479645e-03 -2.616268e-02 -1.479645e-03
   3.712809e-02  5.969998e-03 -1.858466e-01  5.969998e-03 -1.210287e-02  5.969998e-03
   1.042753e-01 -3.533928e-02  8.666140e-01 -3.533928e-02  1.754822e-01 -3.533928e-02
   4.654388e-06 -3.400834e-06  7.566639e-05 -3.400834e-06 -2.118735e-05 -3.400834e-06
   3.910451e-02 -7.995427e-03  1.254902e-01 -7.995427e-03 -3.713318e-02 -7.995427e-03
  -4.208512e-01 -1.331573e-02 -2.938254e-01 -1.331573e-02 -7.662415e-02 -1.331573e-02
   1.085705e-01  3.035147e-02  7.779273e-01  3.035147e-02  1.519828e-02  3.035147e-02
   2.287510e-02 -8.296763e-03  4.540896e-01 -8.296763e-03 -3.442227e-02 -8.296763e-03
   2.215770e-02 -1.417913e-02  9.626192e-02 -1.417912e-02  2.688760e-01 -1.417912e-02
   9.680274e-04  2.038789e-02  8.345274e-01  2.038789e-02 -9.665322e-02  2.038789e-02
   3.966300e-02  2.156311e-02  7.922995e-01  2.156311e-02  1.182360e-02  2.156311e-02
  -5.690763e-02 -5.216437e-03 -1.716680e-01 -5.216437e-03  4.375077e-03 -5.216437e-03
   3.041693e-05  1.264864e-06  5.827101e-05  1.264864e-06 -2.239107e-05  1.264864e-06
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]
   3.034540e-02  1.854930e-03  1.726933e-03  1.854930e-03 -6.492316e-03  1.854930e-03
   5.233935e-02 -7.379720e-03 -3.604561e-02 -7.379720e-03  4.032329e-02 -7.379719e-03
   2.590630e-07 -2.292569e-08 -3.721626e-06 -2.292569e-08 -5.367256e-08 -2.292569e-08
   8.230710e-03 -1.337555e-03  2.459634e-02 -1.337555e-03 -1.828582e-02 -1.337555e-03
  -1.859498e-02  1.386181e-03 -6.034788e-02  1.386181e-03  4.872964e-02  1.386181e-03
  -7.125620e-02 -3.775344e-05 -3.660959e-03 -3.775343e-05  3.554889e-02 -3.775343e-05
   1.963645e-02  5.601187e-03  1.694717e-02  5.601187e-03  2.547934e-02  5.601187e-03
   5.699814e-02 -3.529290e-03  2.362316e-02 -3.529290e-03 -2.874923e-03 -3.529290e-03
   1.073117e-02  1.761930e-04 -1.519782e-02  1.761930e-04 -2.392031e-02  1.761930e-04
  -6.156893e-02 -2.644516e-03 -9.967163e-03 -2.644516e-03  1.653575e-02 -2.644516e-03
   7.335508e-02 -7.585769e-03 -5.588337e-02 -7.585769e-03  2.233629e-02 -7.585769e-03
   4.052034e-02 -1.243587e-02 -9.023810e-03 -1.243587e-02  4.874589e-03 -1.243587e-02
   4.255857e-01  1.660893e-01  1.527786e+00  1.660893e-01  5.655430e-01  1.660893e-01
  -4.068997e-06  1.443652e-06  2.084108e-05  1.443652e-06 -9.354291e-07  1.443652e-06
  -9.837447e-02  5.669977e-02  1.165786e-01  5.669977e-02  1.029714e-01  5.669977e-02
   8.932004e-01 -7.371447e-02 -8.956084e-02 -7.371447e-02 -1.842095e-01 -7.371447e-02
   3.830638e-02 -4.039447e-02  3.403796e-01 -4.039447e-02  2.309060e-01 -4.039447e-02
   4.945708e-02 -1.093331e-02  4.745003e-01 -1.093331e-02  1.906744e-01 -1.093331e-02
   2.048780e+00  1.914477e-01 -5.244093e-02  1.914477e-01  4.951434e-01  1.914477e-01
   2.072210e-01 -8.054736e-02  1.056413e+00 -8.054736e-02  1.384455e+00 -8.054736e-02
   1.011093e-01 -8.843048e-02  1.408717e+00 -8.843048e-02  1.935702e-01 -8.843048e-02
   8.572205e-01  1.724803e-02  4.018319e-02  1.724803e-02  2.227686e-01  1.724803e-02
   1.931595e-06 -1.880187e-05  8.405210e-05 -1.880187e-05  1.480453e-04 -1.880187e-05
          [,43]
  -3.584358e+00
  -4.979314e-01
  -3.995173e-06
  -4.224295e+00
  -4.992701e+00
  -4.430823e+00
  -1.778852e+00
  -3.594938e+00
  -1.890926e+00
  -2.465184e+00
  -2.888071e+00
  -2.770857e+00
  -7.420494e+01
  -1.405915e-03
  -5.567671e+00
  -6.784043e+01
  -2.488542e+01
  -1.469753e+01
  -6.996221e+01
  -5.483706e+01
  -1.749645e+01
  -4.127840e+01
  -6.502722e-03
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  10181 / 10181 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value
s(measurement.no)                            9.00e+00 8.17e+00    0.99    0.20
s(f0Zscore2)                                 9.00e+00 2.00e+00    0.99    0.18
s(durationZscore2)                           2.90e+01 1.00e+00    1.07    1.00
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 2.83e+01    1.00    0.39
ti(measurement.no,durationZscore2)           4.40e+01 1.17e+01    1.01    0.73
ti(measurement.no,f0Zscore2)                 2.80e+01 9.71e+00    1.01    0.88
ti(f0Zscore2,durationZscore2)                7.70e+01 1.23e+01    1.02    0.94
s(word)                                      3.10e+02 1.48e+02      NA      NA
s(wordPos)                                   4.62e+02 2.92e-03      NA      NA
s(wordLeftRightTone)                         4.22e+02 1.11e+01      NA      NA
s(measurement.no,wordPos)                    4.62e+02 1.36e+02      NA      NA
s(f0Zscore2,wordPos)                         4.62e+02 4.98e+01      NA      NA
s(durationZscore2,wordPos)                   4.62e+02 2.94e+01      NA      NA
s(measurement.no,wordLeftRightTone)          4.22e+02 1.40e+02      NA      NA
s(f0Zscore2,wordLeftRightTone)               4.22e+02 1.10e+02      NA      NA
s(durationZscore2,wordLeftRightTone)         4.22e+02 3.50e+01      NA      NA
s(measurement.no,word)                       3.10e+02 8.26e+01      NA      NA
s(f0Zscore2,word)                            3.10e+02 1.31e-02      NA      NA
s(durationZscore2,word)                      3.10e+02 1.47e-03      NA      NA
s(measurement.no,speaker)                    1.00e+02 5.13e+01    0.99    0.20
s(durationZscore2,speaker)                   3.00e+02 2.03e+01    1.07    1.00
s(f0Zscore2,speaker)                         1.00e+02 4.41e+01    0.99    0.18
s(measurement.no,speakerPos)                 3.00e+02 5.14e+01    0.99    0.21
s(durationZscore2,speakerPos)                9.00e+02 8.25e+01    1.07    1.00
s(f0Zscore2,speakerPos)                      3.00e+02 3.78e+01    0.99    0.20
s(measurement.no,speakerLeftRightTone)       5.80e+02 1.07e+02    0.99    0.21
s(durationZscore2,speakerLeftRightTone)      1.74e+03 9.60e+01    1.07    1.00
s(f0Zscore2,speakerLeftRightTone)            5.80e+02 8.20e+01    0.99    0.17

duration0.25auf = quantile(data.au.fem$durationZscore2, c(0.25)) * global_sddauf + global_meandauf

duration0.5auf = quantile(data.au.fem$durationZscore2, c(0.5)) * global_sddauf + global_meandauf

duration0.75auf = quantile(data.au.fem$durationZscore2, c(0.75)) * global_sddauf + global_meandauf

# 3D plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.65512883823603. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.135002585755248. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.536034051323852. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2d, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.au.fem$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.au.fem$durationZscore2, c(0.05,0.95), na.rm = T),
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddauf + global_meandauf), 
           #zlim = c(-0.5,0.5),
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 30 values ranging from -1.395442 to 1.808743. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)

#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")

# 3D plotting
 png("pred2-10.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.65512883823603. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-11.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.135002585755248. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-12.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2d, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf),
        transform = function(f1Zscore2) f1Zscore2 * global_sd1auf + global_mean1auf, 
        zlim = c(500,900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.536034051323852. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(500,900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2-diff-4.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
plot_diff2(gamm.model2d, view=c("measurement.no","durationZscore2"),
           main = "",
           comp = list(f0Zscore2 = quantile(data.au.fem$f0Zscore2, c(0.8,0.2), na.rm = T)),
           ylim = quantile(data.au.fem$durationZscore2, c(0.05,0.95), na.rm = T),
           n.grid = 100,
           transform.view = c(function(measurement.no) measurement.no, function(durationZscore2) durationZscore2 * global_sddauf + global_meandauf), 
           print.summary = TRUE,
           sim.ci = F,
           show.diff = F,
           #color = "heat",
           alpha.diff = 0.5,
           #plot.type = "perp",
           #n.grid =300,
         add.color.legend = FALSE, rm.ranef = TRUE,  
        xlab = "Time (Normalized)", ylab = "Duration (ms)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, hide.label = T)

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * durationZscore2 : numeric predictor; with 100 values ranging from -1.395442 to 1.808743. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

#gradientLegend(valRange=c(-50,20), length=.5, pos=.75, depth = 0.02, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols)
axis(1, at=tickvals, labels=ticknames, las=1, cex.axis=1)
#abline(h = (1.2 * global_sdd1 + global_meand1), lty=2,lwd=2, col = "white")
dev.off()

null device 
          1 

5.7 Model 2E: /ai/ male, F2 as output

system.time(gamm.model2e.noAR <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2e.noAR, paste("Gamm_model2e_noAR.rds"))

gamm.model2e.noAR <- 
  readRDS("Gamm_model2e_noAR.rds")

r.gamm.model2e <- start_value_rho(gamm.model2e.noAR)

system.time(gamm.model2e <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
                      
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2e, AR.start = data.ai.mas$start)
)

#saveRDS(gamm.model2e, paste("Gamm_model2e.rds"))

gamm.model2e <- 
  readRDS("Gamm_model2e.rds")

summary(gamm.model2e, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.03309    0.06372  -0.519    0.603

Approximate significance of smooth terms:
                                                edf Ref.df      F p-value    
s(measurement.no)                             6.587  7.036 42.221 < 2e-16 ***
s(f0Zscore2)                                  1.001  1.001  0.704 0.40177    
s(durationZscore2)                            1.000  1.000  0.070 0.79155    
ti(measurement.no,f0Zscore2,durationZscore2) 43.112 63.333  1.502 0.00624 ** 
ti(measurement.no,durationZscore2)            8.124 10.520  8.408 < 2e-16 ***
ti(measurement.no,f0Zscore2)                  9.596 12.833  1.324 0.18437    
ti(f0Zscore2,durationZscore2)                 4.004  5.519  0.724 0.58084    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.762   Deviance explained = 79.6%
fREML = 7659.4  Scale est. = 0.1865    n = 11029

gam.check(gamm.model2e)

Method: fREML   Optimizer: perf chol
$grad
 [1] -6.466605e-11 -1.099589e-04 -8.205627e-05 -9.580248e-11 -5.427037e-11  2.728839e-11 -3.064549e-11 -5.168244e-11  7.965251e-11
[10] -1.480534e-09 -2.779021e-11 -2.173628e-11  2.299601e-10 -1.102833e-10  2.218954e-10 -3.369180e-10 -1.183778e-09  1.015632e-12
[19] -7.514487e-10 -1.523663e-09  7.680612e-11 -8.705854e-10 -7.873306e-10 -6.887912e-11  3.979395e-11 -4.629656e-05  2.742766e-10
[28] -4.629656e-05 -6.677240e-05 -4.629656e-05 -1.087983e-10 -8.439471e-12  2.441283e-10 -8.387957e-12 -6.500755e-11 -8.248069e-12
[37] -1.099849e-10 -5.539125e-12  2.314096e-10 -5.528911e-12 -1.287361e-10 -5.548895e-12  1.035460e-08

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   2.499957e+00 -1.856674e-06 -1.653204e-08 -1.508445e-02 -1.672591e-03  4.329405e-03  4.331140e-03  4.650016e-03 -6.754962e-02
  -1.856674e-06  1.099444e-04  1.020838e-11  1.393029e-05  3.427163e-06  3.938975e-06 -4.040920e-06  1.819914e-06  3.413605e-06
  -1.653204e-08  1.020838e-11  8.204896e-05 -1.815967e-07 -2.755302e-07 -2.001405e-07  8.574229e-08 -8.465774e-07 -3.572615e-08
  -1.508445e-02  1.393029e-05 -1.815967e-07  2.017139e+00  2.915528e-01  2.097776e-01  1.252796e-04 -6.317605e-02 -4.054230e-02
  -1.672591e-03  3.427163e-06 -2.755302e-07  2.915528e-01  1.362692e+00 -2.210867e-02  6.593857e-03  4.478569e-03 -3.364210e-02
   4.329405e-03  3.938975e-06 -2.001405e-07  2.097776e-01 -2.210867e-02  1.227583e+00 -5.263585e-03 -5.290433e-02  1.008190e-02
   4.331140e-03 -4.040920e-06  8.574229e-08  1.252796e-04  6.593857e-03 -5.263585e-03  8.333206e-01  2.867781e-01 -3.340592e-02
   4.650016e-03  1.819914e-06 -8.465774e-07 -6.317605e-02  4.478569e-03 -5.290433e-02  2.867781e-01  7.909208e-01 -1.490027e-02
  -6.754962e-02  3.413605e-06 -3.572615e-08 -4.054230e-02 -3.364210e-02  1.008190e-02 -3.340592e-02 -1.490027e-02  6.294332e-01
  -2.746060e-02 -6.276876e-05  6.140013e-08  4.948368e-03 -3.964251e-02 -8.079063e-03  9.786722e-05  1.553199e-02  4.819707e-02
   3.353791e-03 -2.192985e-06  2.788012e-08 -7.919451e-03 -1.590065e-02  4.658552e-03 -1.651135e-02  3.210911e-02  6.092042e-04
   1.429716e-04 -6.574118e-07  1.851162e-07  4.083438e-03 -1.125033e-04 -1.252218e-02 -1.212269e-02  9.956261e-02  3.256245e-03
   1.694969e-02  3.095005e-05 -4.186345e-07 -8.227189e-03 -2.994921e-02 -2.731581e-02 -2.615180e-04  5.693877e-02 -1.983557e-02
   5.032234e-03  1.543397e-05 -4.960672e-07 -8.758516e-03  1.935703e-02  3.199550e-02  5.345708e-03  3.775872e-02 -7.938288e-03
  -5.214725e-03  1.899218e-05 -8.057912e-07 -3.949872e-02 -4.525950e-02 -2.620752e-02  1.215210e-02  2.274698e-02 -3.310704e-02
  -6.913185e-03  1.691916e-05  1.296798e-07 -1.682020e-02 -1.317589e-01 -1.532159e-01 -3.759230e-02 -7.200230e-02 -5.290230e-03
   1.378823e-03  2.730559e-07  1.393753e-07  6.044157e-03  2.064937e-02  2.916193e-02  8.434014e-03  6.628455e-02  5.072443e-04
   5.526265e-04 -6.456164e-08 -2.763226e-07  1.337846e-03  1.023882e-03  2.879178e-03 -1.877102e-03 -3.381944e-03  1.018394e-03
  -2.965836e-03 -6.754807e-06 -6.707781e-07  4.890695e-02  4.617717e-02 -8.166221e-02 -1.426889e-02  2.553092e-02  2.861862e-03
   4.854050e-03 -2.822043e-05 -6.249795e-08 -3.139520e-03 -8.841150e-03 -7.316079e-03 -9.275157e-04  2.923955e-02 -3.600278e-03
   5.816375e-04  2.658009e-06 -4.966738e-07 -1.814490e-02 -3.599721e-02  9.226216e-03 -2.901920e-03  8.961434e-03  3.555450e-03
   1.355703e-03  3.064987e-05 -1.731879e-07  3.849314e-02  7.538344e-02  4.284783e-03 -1.920635e-02 -2.025780e-03 -4.753179e-03
   1.187339e-03 -5.531510e-06 -3.605175e-08  3.622007e-03  7.054614e-03 -1.193543e-03  3.086482e-04  9.034116e-03  2.125985e-03
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -2.746060e-02  3.353791e-03  1.429716e-04  1.694969e-02  5.032234e-03 -5.214725e-03 -6.913185e-03  1.378823e-03  5.526265e-04
  -6.276876e-05 -2.192985e-06 -6.574118e-07  3.095005e-05  1.543397e-05  1.899218e-05  1.691916e-05  2.730559e-07 -6.456164e-08
   6.140013e-08  2.788012e-08  1.851162e-07 -4.186345e-07 -4.960672e-07 -8.057912e-07  1.296798e-07  1.393753e-07 -2.763226e-07
   4.948368e-03 -7.919451e-03  4.083438e-03 -8.227189e-03 -8.758516e-03 -3.949872e-02 -1.682020e-02  6.044157e-03  1.337846e-03
  -3.964251e-02 -1.590065e-02 -1.125033e-04 -2.994921e-02  1.935703e-02 -4.525950e-02 -1.317589e-01  2.064937e-02  1.023882e-03
  -8.079063e-03  4.658552e-03 -1.252218e-02 -2.731581e-02  3.199550e-02 -2.620752e-02 -1.532159e-01  2.916193e-02  2.879178e-03
   9.786722e-05 -1.651135e-02 -1.212269e-02 -2.615180e-04  5.345708e-03  1.215210e-02 -3.759230e-02  8.434014e-03 -1.877102e-03
   1.553199e-02  3.210911e-02  9.956261e-02  5.693877e-02  3.775872e-02  2.274698e-02 -7.200230e-02  6.628455e-02 -3.381944e-03
   4.819707e-02  6.092042e-04  3.256245e-03 -1.983557e-02 -7.938288e-03 -3.310704e-02 -5.290230e-03  5.072443e-04  1.018394e-03
   7.703904e-01  3.379696e-03 -2.868638e-03 -1.700091e-02  4.597510e-02 -1.203136e-02 -1.597394e-02  7.573611e-02  1.006841e-03
   3.379696e-03  2.282145e-01 -3.494891e-02  1.889486e-02 -4.540753e-03  2.171057e-02 -2.642886e-02  1.666895e-02  3.606706e-04
  -2.868638e-03 -3.494891e-02  1.823850e-01  7.550811e-04 -3.243549e-02 -1.284960e-03 -1.164307e-02 -1.637415e-02 -2.367336e-03
  -1.700091e-02  1.889486e-02  7.550811e-04  1.176055e+01  2.893151e+00  6.447481e+00 -3.067810e-01 -1.751399e-01  3.898712e-02
   4.597510e-02 -4.540753e-03 -3.243549e-02  2.893151e+00  6.383884e+00  2.325093e+00 -5.542282e-01 -3.873579e-01  7.410274e-02
  -1.203136e-02  2.171057e-02 -1.284960e-03  6.447481e+00  2.325093e+00  1.048642e+01 -2.555864e-01  3.753112e-01  8.773066e-02
  -1.597394e-02 -2.642886e-02 -1.164307e-02 -3.067810e-01 -5.542282e-01 -2.555864e-01  1.612029e+01  6.738540e-01  2.298189e-04
   7.573611e-02  1.666895e-02 -1.637415e-02 -1.751399e-01 -3.873579e-01  3.753112e-01  6.738540e-01  6.509617e+00  6.212265e-03
   1.006841e-03  3.606706e-04 -2.367336e-03  3.898712e-02  7.410274e-02  8.773066e-02  2.298189e-04  6.212265e-03  1.066605e-01
   9.728462e-04  1.569699e-02  9.885661e-03 -1.049705e+00 -2.240141e-01 -1.419817e+00  4.547464e+00  8.984995e-01 -3.152354e-02
   2.969544e-02 -3.644373e-02 -2.764809e-02 -1.850446e-01  2.635104e-01 -2.895171e-01  4.792072e-01  1.518106e+00  4.063479e-04
   4.483759e-02  6.976701e-03 -7.968751e-03  9.682716e-01  5.523419e-01  1.327565e+00 -1.701616e-01  1.643012e-01  3.321198e-01
   2.531893e-02 -5.890767e-03 -1.981867e-02 -2.414355e+00 -4.426306e-01 -1.365744e+00  6.367220e+00  7.546993e-01 -9.244719e-03
   7.922852e-03 -2.470322e-03 -4.434760e-03 -8.462770e-02  2.198132e-02 -6.749337e-02  9.126940e-02  2.848854e-01 -6.796088e-04
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
  -2.965836e-03  4.854050e-03  5.816375e-04  1.355703e-03  1.187339e-03 -4.113183e-04  1.179468e-01 -9.216466e-09  2.800647e-03
  -6.754807e-06 -2.822043e-05  2.658009e-06  3.064987e-05 -5.531510e-06  1.114552e-05 -4.198774e-06 -4.922388e-11  2.162462e-05
  -6.707781e-07 -6.249795e-08 -4.966738e-07 -1.731879e-07 -3.605175e-08 -4.551041e-07 -5.725667e-08  1.461538e-12 -4.703093e-05
   4.890695e-02 -3.139520e-03 -1.814490e-02  3.849314e-02  3.622007e-03 -6.336813e-03  1.588425e-02 -2.985631e-08 -1.456310e-02
   4.617717e-02 -8.841150e-03 -3.599721e-02  7.538344e-02  7.054614e-03 -9.779795e-03  3.648288e-02 -9.442748e-07  2.713929e-02
  -8.166221e-02 -7.316079e-03  9.226216e-03  4.284783e-03 -1.193543e-03  2.021668e-02  2.152976e-02 -2.573275e-07  3.946902e-02
  -1.426889e-02 -9.275157e-04 -2.901920e-03 -1.920635e-02  3.086482e-04 -2.157306e-03 -3.405775e-02  1.433783e-07 -1.076043e-03
   2.553092e-02  2.923955e-02  8.961434e-03 -2.025780e-03  9.034116e-03  2.782671e-03 -1.725456e-02  1.321984e-07 -3.254683e-02
   2.861862e-03 -3.600278e-03  3.555450e-03 -4.753179e-03  2.125985e-03 -9.964253e-04  3.418556e-02  1.081611e-08 -6.979799e-03
   9.728462e-04  2.969544e-02  4.483759e-02  2.531893e-02  7.922852e-03  9.094830e-03  2.656654e-02 -4.111322e-07  5.350468e-02
   1.569699e-02 -3.644373e-02  6.976701e-03 -5.890767e-03 -2.470322e-03  5.758473e-03  1.252263e-02 -1.627739e-07 -7.337454e-04
   9.885661e-03 -2.764809e-02 -7.968751e-03 -1.981867e-02 -4.434760e-03  1.176595e-02  7.202086e-03  5.162879e-07 -1.369324e-02
  -1.049705e+00 -1.850446e-01  9.682716e-01 -2.414355e+00 -8.462770e-02  3.820800e-01 -3.589259e-02  3.066801e-06 -9.840155e-02
  -2.240141e-01  2.635104e-01  5.523419e-01 -4.426306e-01  2.198132e-02  1.925209e-01  7.960997e-03 -4.887124e-06  2.754215e-01
  -1.419817e+00 -2.895171e-01  1.327565e+00 -1.365744e+00 -6.749337e-02  3.957605e-01 -2.796403e-02 -5.532141e-06 -6.764319e-01
   4.547464e+00  4.792072e-01 -1.701616e-01  6.367220e+00  9.126940e-02 -6.887171e-02 -1.079157e-01 -1.669948e-07 -4.633759e-04
   8.984995e-01  1.518106e+00  1.643012e-01  7.546993e-01  2.848854e-01  3.398392e-02  4.603974e-02  2.568205e-06 -3.958268e-01
  -3.152354e-02  4.063479e-04  3.321198e-01 -9.244719e-03 -6.796088e-04  1.261166e-01  3.436145e-03  1.335572e-06 -2.048565e-02
   2.015258e+01  5.381509e-01 -6.705676e-02  1.032069e+01  1.461645e-01  3.839646e-02  9.593669e-02 -2.622182e-06 -1.804629e-01
   5.381509e-01  6.691412e+00 -2.040013e-01  9.735683e-01  3.976005e-01  3.186987e-02  4.370745e-02  5.159481e-06 -3.477845e-01
  -6.705676e-02 -2.040013e-01  5.841671e+00 -2.685912e-01 -2.157379e-02  1.675604e+00 -2.941483e-02 -5.771912e-06  7.393674e-02
   1.032069e+01  9.735683e-01 -2.685912e-01  2.197759e+01  1.639967e-01 -1.187732e-01 -7.048013e-02 -5.741389e-06 -9.436945e-03
   1.461645e-01  3.976005e-01 -2.157379e-02  1.639967e-01  6.879833e-02 -1.976820e-02  8.858230e-03  7.976469e-07 -5.537464e-02
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -9.216466e-09 -9.006027e-07 -9.216466e-09  1.299888e-02 -1.274186e-03 -6.974913e-03 -1.274186e-03 -9.110549e-03 -1.274186e-03
  -4.922388e-11  8.084504e-09 -4.922388e-11 -3.472017e-05  4.427124e-07  3.429839e-05  4.427124e-07  1.552924e-05  4.427124e-07
   1.461538e-12  2.661908e-12  1.461538e-12 -4.532750e-08 -1.071649e-07 -7.178954e-06 -1.071649e-07 -1.058516e-07 -1.071649e-07
  -2.985631e-08  2.877597e-06 -2.985631e-08  1.814238e-01  1.409829e-03  4.507227e-02  1.409829e-03  5.968943e-03  1.409829e-03
  -9.442748e-07  2.611616e-06 -9.442748e-07  8.798849e-03 -3.288914e-03  7.475304e-02 -3.288914e-03 -1.551687e-02 -3.288914e-03
  -2.573275e-07  4.426154e-06 -2.573275e-07  4.079009e-02  4.386275e-03  3.081703e-02  4.386275e-03 -6.150698e-02  4.386275e-03
   1.433783e-07 -7.931023e-07  1.433783e-07 -2.918637e-02 -2.913713e-04 -4.001222e-03 -2.913713e-04 -1.565574e-02 -2.913713e-04
   1.321984e-07 -8.861353e-07  1.321984e-07 -6.168220e-02 -1.748836e-03 -5.455243e-03 -1.748836e-03  1.191002e-02 -1.748836e-03
   1.081611e-08 -1.930117e-06  1.081611e-08 -1.518104e-01  2.577302e-04 -1.277058e-02  2.577302e-04  1.175876e-02  2.577302e-04
  -4.111322e-07 -2.322118e-05 -4.111322e-07  4.380061e-02 -6.757586e-04  6.210171e-02 -6.757586e-04 -1.254586e-01 -6.757586e-04
  -1.627739e-07 -1.238859e-06 -1.627739e-07 -8.847527e-03  4.274908e-04  2.358910e-02  4.274908e-04 -3.278009e-02  4.274908e-04
   5.162879e-07  3.422432e-06  5.162879e-07 -1.261428e-02 -1.805992e-03 -2.896514e-02 -1.805992e-03  3.293708e-02 -1.805992e-03
   3.066801e-06  1.600609e-05  3.066801e-06 -6.049806e-02 -4.458479e-02  6.073267e-01 -4.458479e-02  6.202832e-02 -4.458479e-02
  -4.887124e-06  1.134016e-05 -4.887124e-06 -7.852471e-02 -1.755001e-01  2.115692e-01 -1.755001e-01  1.616274e-02 -1.755001e-01
  -5.532141e-06  3.954329e-05 -5.532141e-06 -4.530259e-03 -4.110957e-02  1.102678e+00 -4.110957e-02  4.341969e-01 -4.110957e-02
  -1.669948e-07  2.277366e-05 -1.669948e-07  1.322945e-01 -3.370702e-02 -4.220903e-01 -3.370702e-02  2.302064e-01 -3.370702e-02
   2.568205e-06  9.832117e-06  2.568205e-06 -6.217306e-02 -2.650030e-02  3.264588e-01 -2.650031e-02  1.251803e-01 -2.650031e-02
   1.335572e-06  2.518505e-06  1.335572e-06  1.172713e-02  2.350061e-02  8.291029e-02  2.350061e-02  4.849956e-02  2.350061e-02
  -2.622182e-06 -1.464153e-05 -2.622182e-06 -4.460914e-01 -4.639843e-02 -1.771732e-01 -4.639843e-02 -2.267621e-01 -4.639843e-02
   5.159481e-06 -1.439540e-05  5.159481e-06  1.435004e-01  4.630446e-02 -5.291383e-01  4.630446e-02 -5.288312e-02  4.630446e-02
  -5.771912e-06  1.050941e-05 -5.771912e-06 -7.510610e-03 -4.545618e-03 -1.449367e-01 -4.545618e-03  3.732697e-01 -4.545618e-03
  -5.741389e-06  1.353912e-05 -5.741389e-06  2.257074e-02 -2.649811e-02 -1.374628e-01 -2.649811e-02  3.958191e-02 -2.649811e-02
   7.976469e-07 -2.506289e-07  7.976469e-07  1.136349e-02  8.714734e-04 -6.151112e-02  8.714734e-04  5.677868e-03  8.714734e-04
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]         [,43]
   7.517359e-02  2.776337e-04 -6.855362e-03  2.776337e-04 -5.089686e-03  2.776337e-04 -2.793525e+00
   3.317663e-07  1.675043e-06  2.341829e-05  1.675043e-06  1.130294e-05  1.675043e-06 -2.912291e-04
   1.014351e-07 -1.833834e-07 -1.480195e-05 -1.833834e-07 -8.691400e-07 -1.833834e-07 -1.088927e-05
  -1.897037e-01  5.312938e-03 -5.373390e-02  5.312938e-03  7.018265e-02  5.312938e-03 -7.328433e+00
   5.623876e-02 -2.768893e-04  1.785247e-02 -2.768893e-04  1.096926e-01 -2.768893e-04 -7.928704e+00
  -1.610249e-02 -1.815318e-03 -5.325442e-02 -1.815318e-03  4.768910e-02 -1.815318e-03 -5.798656e+00
  -4.172417e-02  1.291744e-03  1.405670e-04  1.291744e-03  2.407981e-03  1.291744e-03 -1.568555e+00
  -2.682772e-02  7.789009e-03 -8.839564e-03  7.789009e-03  6.169620e-02  7.789009e-03 -1.993330e+00
  -8.278395e-03 -1.615842e-03 -6.759531e-02 -1.615842e-03 -3.107265e-02 -1.615842e-03 -1.753660e+00
  -2.052793e-02  5.569134e-03  4.839447e-02  5.569134e-03 -1.322024e-01  5.569134e-03 -2.544212e+00
  -3.585578e-02 -5.556687e-03  1.618409e-02 -5.556687e-03  4.154559e-02 -5.556687e-03 -7.021839e-01
  -1.134391e-02 -4.965415e-03 -1.819228e-02 -4.965415e-03  6.003034e-02 -4.965415e-03 -7.998985e-01
   2.027300e-01 -4.465192e-03 -4.038961e-01 -4.465192e-03  2.733135e-02 -4.465192e-03 -3.200164e+01
  -7.797114e-02  3.589740e-02  1.388210e-01  3.589740e-02  1.918560e-01  3.589740e-02 -2.324852e+01
   9.813057e-02 -2.262473e-02 -9.669716e-01 -2.262473e-02 -8.764547e-02 -2.262473e-02 -3.533263e+01
  -3.364522e-01 -3.115238e-02  1.712298e-01 -3.115238e-02  8.236267e-02 -3.115238e-02 -4.425150e+01
   3.036085e-01  2.073216e-02  6.482705e-02  2.073216e-02  1.304600e-01  2.073216e-02 -2.586444e+01
   1.147328e-02  1.004829e-02  1.573968e-01  1.004829e-02  2.991350e-03  1.004829e-02 -2.044627e+00
   2.486090e-01 -3.430846e-02 -2.822484e-01 -3.430846e-02 -1.485171e-01 -3.430846e-02 -4.964285e+01
  -3.092739e-01  1.370718e-02  2.567141e-01  1.370718e-02  3.114682e-01  1.370718e-02 -2.372548e+01
   8.471697e-02  1.515179e-02  1.790485e+00  1.515179e-02  5.428370e-02  1.515179e-02 -2.202967e+01
   2.816684e-01 -1.389207e-02 -4.396930e-01 -1.389207e-02 -2.355564e-02 -1.389207e-02 -4.977567e+01
   2.006597e-02 -7.143762e-04 -3.748551e-02 -7.143761e-04  2.813589e-02 -7.143761e-04 -2.466024e+00
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  8855 / 8855 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value    
s(measurement.no)                            9.00e+00 6.59e+00    0.99   0.305    
s(f0Zscore2)                                 9.00e+00 1.00e+00    1.00   0.545    
s(durationZscore2)                           2.90e+01 1.00e+00    0.99   0.145    
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 4.31e+01    0.95  <2e-16 ***
ti(measurement.no,durationZscore2)           4.40e+01 8.12e+00    1.03   0.970    
ti(measurement.no,f0Zscore2)                 2.80e+01 9.60e+00    1.01   0.675    
ti(f0Zscore2,durationZscore2)                7.70e+01 4.00e+00    0.98   0.065 .  
s(word)                                      2.09e+02 6.40e+01      NA      NA    
s(wordPos)                                   3.31e+02 4.65e+01      NA      NA    
s(wordLeftRightTone)                         3.10e+02 7.07e+01      NA      NA    
s(measurement.no,wordPos)                    3.31e+02 8.85e+01      NA      NA    
s(f0Zscore2,wordPos)                         3.31e+02 5.17e+01      NA      NA    
s(durationZscore2,wordPos)                   3.31e+02 4.09e+00      NA      NA    
s(measurement.no,wordLeftRightTone)          3.10e+02 9.93e+01      NA      NA    
s(f0Zscore2,wordLeftRightTone)               3.10e+02 4.75e+01      NA      NA    
s(durationZscore2,wordLeftRightTone)         3.10e+02 4.41e+01      NA      NA    
s(measurement.no,word)                       2.09e+02 9.96e+01      NA      NA    
s(f0Zscore2,word)                            2.09e+02 4.93e+00      NA      NA    
s(durationZscore2,word)                      2.09e+02 1.42e+01      NA      NA    
s(measurement.no,speaker)                    1.00e+02 3.84e+01    0.99   0.285    
s(durationZscore2,speaker)                   3.00e+02 1.04e+02    0.99   0.155    
s(f0Zscore2,speaker)                         1.00e+02 4.19e-03    1.00   0.505    
s(measurement.no,speakerPos)                 3.00e+02 7.37e+01    0.99   0.320    
s(durationZscore2,speakerPos)                9.00e+02 1.05e+02    0.99   0.195    
s(f0Zscore2,speakerPos)                      3.00e+02 6.74e+01    1.00   0.515    
s(measurement.no,speakerLeftRightTone)       5.90e+02 1.17e+02    0.99   0.245    
s(durationZscore2,speakerLeftRightTone)      1.77e+03 2.59e+02    0.99   0.100 .  
s(f0Zscore2,speakerLeftRightTone)            5.90e+02 1.11e+02    1.00   0.580    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# Plotting
# 3D plot 
# png("pred2b-1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
# dev.off()
# png("pred2b-2.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
# dev.off()
# png("pred2b-3.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

# dev.off()

# Plotting
# 3D plot 
png("pred2b-1.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.449662795901817. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-2.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.0671853722597352. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-3.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2e, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75 * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0 + global_mean0), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2 + global_mean2, 
        zlim = c(1400,1900),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.266227 to 1.296824. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.637773887350503. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Deb. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 H H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0011. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0006 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0002 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1400,1900), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

5.8 Model 2F: /ai/ female, F2 as output

system.time(gamm.model2f.noAR <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2f.noAR, paste("Gamm_model2f_noAR.rds"))

gamm.model2f.noAR <- 
  readRDS("Gamm_model2f_noAR.rds")

r.gamm.model2f <- start_value_rho(gamm.model2f.noAR)

system.time(gamm.model2f <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
                      
                      data=data.ai.fem, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2f, AR.start = data.ai.fem$start)
)

#saveRDS(gamm.model2f, paste("Gamm_model2f.rds"))

gamm.model2f <- 
  readRDS("Gamm_model2f.rds")

summary(gamm.model2f, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.001353   0.068175    0.02    0.984

Approximate significance of smooth terms:
                                                edf Ref.df      F  p-value    
s(measurement.no)                             1.000  1.000 43.012  < 2e-16 ***
s(f0Zscore2)                                  1.299  1.383  1.414   0.3345    
s(durationZscore2)                            1.000  1.000  1.746   0.1864    
ti(measurement.no,f0Zscore2,durationZscore2)  7.153 11.575  0.330   0.9841    
ti(measurement.no,durationZscore2)           11.826 15.816  3.599 1.74e-06 ***
ti(measurement.no,f0Zscore2)                  6.133  8.041  2.059   0.0377 *  
ti(f0Zscore2,durationZscore2)                 1.990  2.644  0.067   0.9769    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.565   Deviance explained = 61.7%
fREML =  10576  Scale est. = 0.39103   n = 10275

gam.check(gamm.model2f)

Method: fREML   Optimizer: perf chol
$grad
 [1] -4.059069e-05 -3.621745e-11 -6.621412e-05 -2.511880e-12  9.776513e-12  1.672240e-11  2.539258e-11 -2.151834e-11  8.259393e-12
[10]  2.543565e-11 -4.782791e-11 -4.194253e-11 -1.131390e-09 -1.109672e-10 -6.864180e-10  3.423395e-11 -2.213056e-10 -8.591172e-10
[19]  1.992291e-10 -7.025869e-05 -8.696881e-05  5.742606e-11 -3.369820e-10 -3.685521e-10 -2.665956e-11 -1.379341e-12 -3.918283e-05
[28] -1.368031e-12  8.075673e-11 -1.376385e-12  5.211120e-11  1.643352e-12  4.817480e-12  1.604161e-12  6.907186e-11  1.598943e-12
[37]  3.115730e-11  3.152856e-11  2.128644e-10  3.161116e-11 -6.675549e-11  3.155920e-11  1.219723e-08

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   4.058776e-05 -6.835414e-08 -9.255582e-13  2.124997e-07  3.148291e-08  1.198916e-07 -3.751491e-07  5.056158e-08 -1.556535e-06
  -6.835414e-08  2.801906e-02  1.807924e-08  1.082223e-03 -1.231766e-03 -7.797321e-04 -2.451471e-03 -3.457879e-03 -1.321340e-02
  -9.255582e-13  1.807924e-08  6.620930e-05  1.722044e-08 -2.723727e-08  1.565426e-08 -5.063545e-08 -9.382181e-08  6.583246e-09
   2.124997e-07  1.082223e-03  1.722044e-08  1.617203e-01 -6.588701e-03 -7.053157e-02 -1.820261e-02  1.523970e-02  6.572580e-04
   3.148291e-08 -1.231766e-03 -2.723727e-08 -6.588701e-03  4.733013e-02 -1.261378e-02  4.434096e-03 -9.960437e-03  1.978154e-03
   1.198916e-07 -7.797321e-04  1.565426e-08 -7.053157e-02 -1.261378e-02  1.346989e-01  4.297510e-03  3.423857e-02  2.442259e-03
  -3.751491e-07 -2.451471e-03 -5.063545e-08 -1.820261e-02  4.434096e-03  4.297510e-03  1.194557e+00  3.119124e-01 -9.322738e-03
   5.056158e-08 -3.457879e-03 -9.382181e-08  1.523970e-02 -9.960437e-03  3.423857e-02  3.119124e-01  1.250384e+00 -1.504501e-03
  -1.556535e-06 -1.321340e-02  6.583246e-09  6.572580e-04  1.978154e-03  2.442259e-03 -9.322738e-03 -1.504501e-03  4.868940e-01
  -4.379806e-06 -1.230206e-02 -5.874370e-08  5.920278e-03 -2.842395e-03  1.214910e-03  4.072969e-03 -7.159884e-03  2.418944e-01
  -6.493615e-08  1.564333e-03  1.502420e-07  1.022127e-03 -3.001984e-03 -1.724736e-03 -3.372672e-03  6.555813e-04 -1.253205e-03
  -1.361317e-08  5.427404e-04  3.029203e-07 -2.461989e-03  1.180844e-03  1.696576e-03  6.963684e-05  7.716564e-03  5.609250e-05
  -1.180111e-07 -2.900070e-03 -1.358259e-06 -1.086932e-03 -5.657394e-03 -2.791764e-04  8.148794e-03  6.568174e-02  1.462764e-02
  -1.617118e-08  2.275393e-03 -1.429857e-07 -6.714172e-04 -1.612705e-04  2.683702e-03 -1.820404e-02 -3.266952e-03  7.660801e-05
   8.847983e-08  3.311472e-04 -9.386240e-08  2.777844e-04 -3.248269e-04  3.041690e-04  2.790342e-04 -7.764825e-03 -7.104994e-04
  -4.817675e-07  7.485082e-03 -1.800745e-07  1.433210e-02 -3.394547e-02 -2.315925e-02  4.784249e-04  1.624133e-01  4.185889e-03
   5.433134e-07 -3.232879e-02  7.058079e-07 -7.211679e-03 -2.501599e-04  4.433364e-03  2.551981e-02 -1.094667e-02  2.127187e-02
  -1.147321e-07 -3.633317e-03 -6.520884e-06  5.540289e-04  5.520110e-03  8.934579e-03 -1.512218e-02  1.676564e-02 -2.613643e-04
  -5.873207e-07  1.450407e-02 -2.362378e-07  1.410898e-02  1.023451e-02  2.017473e-03  1.127844e-02  4.474944e-02 -5.739211e-03
  -7.685524e-12  5.768272e-07 -5.237494e-11 -5.825577e-07  1.029048e-06  1.552167e-07 -1.126665e-06 -5.372108e-06  4.236203e-07
  -5.853066e-11 -6.104601e-07 -1.460963e-09 -1.830255e-06  1.026413e-06 -4.958978e-07 -6.740442e-06  4.303934e-06 -1.619065e-07
  -4.538810e-07  3.648839e-03 -1.607981e-06 -1.112204e-03 -2.281638e-02 -8.067152e-03 -2.225024e-02  6.111324e-02 -4.203095e-03
  -1.852631e-07 -1.378429e-02  7.145217e-07  5.412647e-04 -1.309286e-02  4.607944e-03  2.018895e-02  7.504466e-03  6.099616e-03
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -4.379806e-06 -6.493615e-08 -1.361317e-08 -1.180111e-07 -1.617118e-08  8.847983e-08 -4.817675e-07  5.433134e-07 -1.147321e-07
  -1.230206e-02  1.564333e-03  5.427404e-04 -2.900070e-03  2.275393e-03  3.311472e-04  7.485082e-03 -3.232879e-02 -3.633317e-03
  -5.874370e-08  1.502420e-07  3.029203e-07 -1.358259e-06 -1.429857e-07 -9.386240e-08 -1.800745e-07  7.058079e-07 -6.520884e-06
   5.920278e-03  1.022127e-03 -2.461989e-03 -1.086932e-03 -6.714172e-04  2.777844e-04  1.433210e-02 -7.211679e-03  5.540289e-04
  -2.842395e-03 -3.001984e-03  1.180844e-03 -5.657394e-03 -1.612705e-04 -3.248269e-04 -3.394547e-02 -2.501599e-04  5.520110e-03
   1.214910e-03 -1.724736e-03  1.696576e-03 -2.791764e-04  2.683702e-03  3.041690e-04 -2.315925e-02  4.433364e-03  8.934579e-03
   4.072969e-03 -3.372672e-03  6.963684e-05  8.148794e-03 -1.820404e-02  2.790342e-04  4.784249e-04  2.551981e-02 -1.512218e-02
  -7.159884e-03  6.555813e-04  7.716564e-03  6.568174e-02 -3.266952e-03 -7.764825e-03  1.624133e-01 -1.094667e-02  1.676564e-02
   2.418944e-01 -1.253205e-03  5.609250e-05  1.462764e-02  7.660801e-05 -7.104994e-04  4.185889e-03  2.127187e-02 -2.613643e-04
   4.087111e-01 -3.579560e-03 -3.454027e-04  7.507283e-02  1.352387e-03 -6.644612e-03  3.590664e-02 -1.812703e-02 -1.175214e-02
  -3.579560e-03  1.044992e-02 -6.487705e-03 -1.320430e-03  2.002429e-03 -7.117255e-04  1.085290e-02  5.740033e-03 -2.205535e-03
  -3.454027e-04 -6.487705e-03  1.162602e-02 -1.067867e-04  4.741298e-03 -4.295141e-04  1.375038e-03  5.506141e-03 -1.087743e-02
   7.507283e-02 -1.320430e-03 -1.067867e-04  1.459165e+01  1.902873e+00  4.431315e-01 -5.705216e-01  9.354388e-01  9.516980e-01
   1.352387e-03  2.002429e-03  4.741298e-03  1.902873e+00  3.268664e+00  3.643153e-02  4.998851e-01 -1.039178e-01  1.715275e-01
  -6.644612e-03 -7.117255e-04 -4.295141e-04  4.431315e-01  3.643153e-02  2.121039e-01  2.261631e-02  5.541849e-02  1.149004e-01
   3.590664e-02  1.085290e-02  1.375038e-03 -5.705216e-01  4.998851e-01  2.261631e-02  2.013791e+01  1.449042e+00 -3.929527e-01
  -1.812703e-02  5.740033e-03  5.506141e-03  9.354388e-01 -1.039178e-01  5.541849e-02  1.449042e+00  9.538980e+00  6.206695e-01
  -1.175214e-02 -2.205535e-03 -1.087743e-02  9.516980e-01  1.715275e-01  1.149004e-01 -3.929527e-01  6.206695e-01  3.296439e+00
   2.774524e-02  1.145524e-02 -1.037996e-03 -8.679371e-01 -2.590166e-01 -1.214385e-01  4.201246e+00  4.972705e-01 -2.236343e-02
   1.481589e-06 -6.388545e-07  2.308774e-08  2.608423e-05 -4.435716e-08  1.268650e-05 -5.343321e-06  2.593156e-05  1.242025e-05
  -3.899809e-06  1.795139e-06 -8.741417e-07  1.643075e-04  5.587662e-05 -1.222296e-05 -4.150983e-06  4.645168e-05  3.270579e-04
   3.695766e-02  2.885622e-03 -1.584840e-03 -1.862341e+00 -2.323828e-01 -4.783230e-02  6.204254e+00  7.180152e-01 -9.428536e-02
  -1.076531e-02  5.540443e-03  4.223108e-03  6.307587e-01  1.541098e-01 -6.366158e-04  1.982266e-01  2.760731e+00  1.881498e-01
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
  -5.873207e-07 -7.685524e-12 -5.853066e-11 -4.538810e-07 -1.852631e-07 -1.995891e-08 -1.200739e-05 -4.723865e-10 -5.557988e-13
   1.450407e-02  5.768272e-07 -6.104601e-07  3.648839e-03 -1.378429e-02  1.651868e-06 -3.651729e-03  1.301933e-05 -2.973608e-08
  -2.362378e-07 -5.237494e-11 -1.460963e-09 -1.607981e-06  7.145217e-07 -5.733993e-07 -2.733974e-08 -3.078659e-09 -2.899777e-10
   1.410898e-02 -5.825577e-07 -1.830255e-06 -1.112204e-03  5.412647e-04 -1.444551e-04 -7.701303e-03  4.063392e-05 -1.958458e-08
   1.023451e-02  1.029048e-06  1.026413e-06 -2.281638e-02 -1.309286e-02 -1.261630e-03  6.714412e-04  1.660672e-04  1.168403e-07
   2.017473e-03  1.552167e-07 -4.958978e-07 -8.067152e-03  4.607944e-03 -3.042817e-04 -1.303073e-03  3.625641e-04  1.808565e-07
   1.127844e-02 -1.126665e-06 -6.740442e-06 -2.225024e-02  2.018895e-02 -5.193177e-04  6.055852e-03  1.833099e-05 -6.298191e-08
   4.474944e-02 -5.372108e-06  4.303934e-06  6.111324e-02  7.504466e-03  3.276579e-03  2.317523e-02 -4.248413e-04  1.021437e-07
  -5.739211e-03  4.236203e-07 -1.619065e-07 -4.203095e-03  6.099616e-03 -6.186873e-05  7.145677e-02  1.060776e-06  6.100304e-08
   2.774524e-02  1.481589e-06 -3.899809e-06  3.695766e-02 -1.076531e-02 -9.311176e-04 -1.042423e-02 -2.421610e-04  6.348219e-08
   1.145524e-02 -6.388545e-07  1.795139e-06  2.885622e-03  5.540443e-03  4.547342e-04 -1.393432e-03  9.763462e-05  1.679740e-08
  -1.037996e-03  2.308774e-08 -8.741417e-07 -1.584840e-03  4.223108e-03 -6.005400e-04 -1.838163e-03  1.571946e-05 -4.297755e-08
  -8.679371e-01  2.608423e-05  1.643075e-04 -1.862341e+00  6.307587e-01  1.064780e-01 -1.972778e-02 -1.625717e-03  4.212569e-06
  -2.590166e-01 -4.435716e-08  5.587662e-05 -2.323828e-01  1.541098e-01  6.631936e-04 -2.363023e-02  2.457792e-03 -2.841081e-07
  -1.214385e-01  1.268650e-05 -1.222296e-05 -4.783230e-02 -6.366158e-04  1.154977e-02  1.461398e-02  1.176743e-04 -1.790083e-07
   4.201246e+00 -5.343321e-06 -4.150983e-06  6.204254e+00  1.982266e-01 -3.222782e-02 -1.521591e-01 -1.747439e-03  4.823040e-07
   4.972705e-01  2.593156e-05  4.645168e-05  7.180152e-01  2.760731e+00  2.517874e-02  1.712137e-01 -7.201126e-03 -1.414438e-06
  -2.236343e-02  1.242025e-05  3.270579e-04 -9.428536e-02  1.881498e-01  2.030834e-01  5.467412e-02 -2.953599e-03  1.754523e-06
   2.033196e+01  2.395834e-05 -1.232484e-05  8.753931e+00  2.736612e-01 -2.682660e-03 -1.136062e-02 -4.688517e-03  2.210305e-06
   2.395834e-05  7.025797e-05 -4.081471e-09 -5.898835e-07  3.023141e-05  3.949858e-07  9.776983e-07 -5.436585e-08 -3.161632e-11
  -1.232484e-05 -4.081471e-09  8.698871e-05 -4.534432e-06  4.142141e-05  5.488032e-05  1.464269e-05  8.385178e-08 -3.757786e-10
   8.753931e+00 -5.898835e-07 -4.534432e-06  1.490760e+01  4.303033e-01 -7.513833e-03 -1.297027e-01 -2.610450e-03 -8.819002e-07
   2.736612e-01  3.023141e-05  4.142141e-05  4.303033e-01  3.258479e+00  4.458894e-03 -2.300952e-02 -2.798848e-03  4.384795e-07
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -4.723865e-10 -7.158549e-08 -4.723865e-10 -1.442855e-06  2.064283e-08 -2.869540e-07  2.064283e-08  1.694545e-08  2.064283e-08
   1.301933e-05  5.634651e-02  1.301933e-05 -8.010721e-03  6.247605e-04  7.383497e-03  6.247605e-04 -5.215962e-05  6.247605e-04
  -3.078659e-09 -1.492172e-08 -3.078659e-09 -3.921845e-07  4.573161e-09 -2.051854e-05  4.573161e-09  6.646263e-07  4.573161e-09
   4.063391e-05 -5.289316e-03  4.063391e-05  1.102660e-02 -2.722202e-04 -2.554748e-02 -2.722202e-04  9.496156e-03 -2.722202e-04
   1.660672e-04 -1.218616e-02  1.660672e-04 -6.439294e-03  2.107252e-04 -2.843185e-02  2.107252e-04 -1.433799e-02  2.107252e-04
   3.625641e-04 -3.699236e-02  3.625641e-04  1.654930e-02  8.312763e-04 -1.382463e-02  8.312763e-04 -1.324787e-02  8.312763e-04
   1.833099e-05 -6.699553e-02  1.833099e-05  2.749962e-02  1.974851e-03  2.400890e-02  1.974851e-03 -4.317661e-03  1.974851e-03
  -4.248413e-04 -7.604221e-02 -4.248413e-04  5.490611e-02  1.245795e-03  4.617678e-02  1.245795e-03 -2.789863e-02  1.245795e-03
   1.060776e-06 -2.703012e-02  1.060776e-06 -3.218281e-02  1.036232e-03  1.030730e-02  1.036232e-03  2.094064e-02  1.036232e-03
  -2.421610e-04  2.356735e-02 -2.421610e-04 -1.057850e-02 -1.141307e-03 -1.175262e-02 -1.141307e-03  2.237277e-02 -1.141307e-03
   9.763462e-05 -3.711875e-03  9.763462e-05 -4.351585e-03  8.267097e-04 -3.795412e-03  8.267097e-04 -1.134956e-02  8.267097e-04
   1.571946e-05 -1.051006e-02  1.571946e-05 -7.628076e-03  5.031705e-04 -1.082069e-02  5.031705e-04 -5.153669e-03  5.031705e-04
  -1.625717e-03 -1.227684e-01 -1.625717e-03 -6.463396e-02  8.638910e-03  4.181867e-01  8.638910e-03  3.022208e-02  8.638909e-03
   2.457792e-03  1.439113e-02  2.457792e-03 -5.883933e-02 -4.743485e-05 -1.329131e-01 -4.743485e-05  4.180930e-02 -4.743485e-05
   1.176743e-04  1.047972e-02  1.176743e-04  3.516354e-03  7.139913e-03  3.719543e-02  7.139913e-03  2.610360e-02  7.139913e-03
  -1.747439e-03  2.097532e-02 -1.747439e-03  1.881322e-01 -2.375697e-02 -2.856285e-01 -2.375697e-02  6.611823e-02 -2.375697e-02
  -7.201126e-03 -1.976255e-02 -7.201126e-03  2.119787e-02 -2.222549e-02 -9.514765e-02 -2.222549e-02 -3.439864e-01 -2.222549e-02
  -2.953599e-03 -8.152549e-02 -2.953599e-03  6.626745e-02  4.461932e-03  2.197801e-01  4.461932e-03  4.600066e-02  4.461932e-03
  -4.688517e-03  3.596685e-01 -4.688517e-03 -1.235922e-01 -2.811537e-02  1.151265e-01 -2.811537e-02  2.064333e-01 -2.811537e-02
  -5.436585e-08  5.619420e-06 -5.436585e-08 -1.567984e-06  5.037461e-07 -1.346804e-05  5.037461e-07  3.579519e-06  5.037461e-07
   8.385178e-08 -9.413876e-06  8.385178e-08  4.472862e-06  7.753363e-07  2.155098e-04  7.753362e-07  4.350311e-05  7.753362e-07
  -2.610450e-03  1.119170e-01 -2.610450e-03  6.304050e-02 -3.183822e-02 -2.193641e-01 -3.183822e-02  1.005136e-02 -3.183822e-02
  -2.798848e-03  5.672466e-02 -2.798848e-03  1.079031e-01 -1.273432e-02 -1.016659e-01 -1.273432e-02  1.027159e-01 -1.273432e-02
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]         [,43]
  -2.355174e-06  3.746502e-08  2.567077e-07  3.746502e-08 -5.047977e-08  3.746502e-08 -1.607589e-05
   8.905210e-03  9.545980e-04  8.272806e-03  9.545980e-04  1.792717e-02  9.545980e-04 -1.496071e-01
   1.582531e-07  1.735046e-08 -2.102696e-05  1.735046e-08  4.700635e-07  1.735046e-08 -1.346904e-05
   1.820813e-02 -8.865944e-05 -5.603346e-03 -8.865944e-05  6.202788e-03 -8.865944e-05 -8.367940e-01
   1.036345e-02  2.014739e-03  2.988353e-02  2.014739e-03  3.726011e-03  2.014739e-03 -9.458253e-01
   1.628755e-02  1.250722e-03  2.714633e-02  1.250722e-03  8.180883e-03  1.250722e-03 -1.293794e+00
  -6.609396e-02  5.814416e-03  4.953492e-02  5.814416e-03  8.360095e-02  5.814416e-03 -2.180152e+00
   2.728893e-02  1.277839e-03  1.063570e-01  1.277839e-03 -8.030345e-02  1.277839e-03 -3.232756e+00
  -3.725881e-02  2.750189e-03  2.566467e-03  2.750189e-03 -1.775884e-02  2.750189e-03 -1.196251e+00
  -1.604958e-02  4.341278e-03  4.589217e-02  4.341278e-03 -1.729495e-02  4.341278e-03 -1.370127e+00
  -5.598287e-05 -5.696725e-04  7.848316e-04 -5.696725e-04  1.743347e-02 -5.696725e-04 -2.512830e-01
  -1.768843e-03  1.058570e-04  1.421523e-02  1.058570e-04  1.830551e-02  1.058570e-04 -2.437146e-01
   1.383733e-01 -1.543944e-02  4.372202e-01 -1.543944e-02 -1.357168e-01 -1.543944e-02 -3.504332e+01
   4.719865e-02  1.457266e-02  1.727387e-01  1.457266e-02 -7.422392e-02  1.457266e-02 -1.313519e+01
  -4.092469e-02  8.190201e-04 -1.554662e-01  8.190201e-04 -5.434497e-03  8.190201e-04 -3.235735e+00
  -1.251299e-01  9.712825e-02  2.496001e-02  9.712825e-02  2.167738e-01  9.712825e-02 -4.640936e+01
  -2.267620e-01 -4.018950e-02 -8.049615e-01 -4.018950e-02  2.460837e-01 -4.018950e-02 -3.537426e+01
   1.346285e-01 -1.511220e-02  3.178864e-01 -1.511220e-02 -1.666556e-01 -1.511220e-02 -1.571325e+01
  -1.912494e-01 -1.766386e-01 -7.311826e-01 -1.766386e-01  2.367257e-01 -1.766386e-01 -4.916784e+01
  -1.584603e-06  4.743486e-07 -1.235287e-05  4.743486e-07  1.746024e-05  4.743486e-07 -5.001692e-04
   5.573328e-05 -5.663561e-06  6.961738e-05 -5.663561e-06  7.516810e-06 -5.663561e-06 -3.279942e-03
  -1.733565e-01 -3.382266e-02 -4.885975e-01 -3.382266e-02  1.222144e-01 -3.382266e-02 -3.980497e+01
   1.482106e-01  1.408333e-02  8.745991e-02  1.408333e-02  2.004322e-01  1.408333e-02 -1.760626e+01
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  8663 / 8663 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value  
s(measurement.no)                            9.00e+00 1.00e+00    1.00   0.390  
s(f0Zscore2)                                 9.00e+00 1.30e+00    0.99   0.285  
s(durationZscore2)                           2.90e+01 1.00e+00    0.99   0.255  
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 7.15e+00    1.01   0.660  
ti(measurement.no,durationZscore2)           4.40e+01 1.18e+01    1.01   0.845  
ti(measurement.no,f0Zscore2)                 2.80e+01 6.13e+00    1.01   0.880  
ti(f0Zscore2,durationZscore2)                7.70e+01 1.99e+00    0.98   0.075 .
s(word)                                      2.03e+02 7.01e+01      NA      NA  
s(wordPos)                                   3.09e+02 2.63e+01      NA      NA  
s(wordLeftRightTone)                         2.90e+02 6.47e+00      NA      NA  
s(measurement.no,wordPos)                    3.09e+02 9.28e+01      NA      NA  
s(f0Zscore2,wordPos)                         3.09e+02 7.07e+01      NA      NA  
s(durationZscore2,wordPos)                   3.09e+02 3.14e+01      NA      NA  
s(measurement.no,wordLeftRightTone)          2.90e+02 9.83e+01      NA      NA  
s(f0Zscore2,wordLeftRightTone)               2.90e+02 1.14e-03      NA      NA  
s(durationZscore2,wordLeftRightTone)         2.90e+02 6.73e-03      NA      NA  
s(measurement.no,word)                       2.03e+02 7.96e+01      NA      NA  
s(f0Zscore2,word)                            2.03e+02 3.52e+01      NA      NA  
s(durationZscore2,word)                      2.03e+02 2.21e+00      NA      NA  
s(measurement.no,speaker)                    1.00e+02 5.00e+01    1.00   0.370  
s(durationZscore2,speaker)                   3.00e+02 1.52e-01    0.99   0.245  
s(f0Zscore2,speaker)                         1.00e+02 3.44e+01    0.99   0.300  
s(measurement.no,speakerPos)                 3.00e+02 4.09e+01    1.00   0.420  
s(durationZscore2,speakerPos)                9.00e+02 1.34e+02    0.99   0.230  
s(f0Zscore2,speakerPos)                      3.00e+02 4.96e+01    0.99   0.260  
s(measurement.no,speakerLeftRightTone)       5.90e+02 7.93e+01    1.00   0.360  
s(durationZscore2,speakerLeftRightTone)      1.77e+03 2.35e+02    0.99   0.210  
s(f0Zscore2,speakerLeftRightTone)            5.90e+02 7.52e+01    0.99   0.235  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# Plotting
# 3D plot 

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.627777955022104. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0941417149129249. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.525044865601105. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

# Plotting
# 3D plot 
png("pred2b-4.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.627777955022104. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-5.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0941417149129249. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-6.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2f, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.ai.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.ai.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75f * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0f + global_mean0f), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2f + global_mean2f, 
        zlim = c(1700,2100),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.243937 to 1.329686. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.525044865601105. 
    * word : factor; set to the value(s): 在. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 在 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 在 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0130. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0130 Med. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0130 H H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1700,2100), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

5.9 Model 2G: /au/ male, F2 as output

system.time(gamm.model2g.noAR <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2g.noAR, paste("Gamm_model2g_noAR.rds"))

gamm.model2g.noAR <- 
  readRDS("Gamm_model2g_noAR.rds")

r.gamm.model2g <- start_value_rho(gamm.model2g.noAR)

# Auto-regressive model

system.time(gamm.model2g <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
 
                        
                    
                      data=data.au.mas, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2g, AR.start = data.au.mas$start)
)

#saveRDS(gamm.model2g, paste("Gamm_model2g.rds"))

gamm.model2g <- 
  readRDS("Gamm_model2g.rds")

summary(gamm.model2g, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.007337   0.065831  -0.111    0.911

Approximate significance of smooth terms:
                                                edf Ref.df      F  p-value    
s(measurement.no)                             6.857  7.407 50.956  < 2e-16 ***
s(f0Zscore2)                                  1.001  1.001  3.625  0.05699 .  
s(durationZscore2)                            1.000  1.000  6.844  0.00891 ** 
ti(measurement.no,f0Zscore2,durationZscore2) 64.111 88.432  1.371  0.01136 *  
ti(measurement.no,durationZscore2)           12.116 14.981  7.821  < 2e-16 ***
ti(measurement.no,f0Zscore2)                 12.808 16.209  3.462 3.87e-06 ***
ti(f0Zscore2,durationZscore2)                14.751 18.368  1.484  0.07181 .  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =  0.768   Deviance explained = 80.9%
fREML = 7928.2  Scale est. = 0.20584   n = 9522

gam.check(gamm.model2g)

Method: fREML   Optimizer: perf chol
$grad
 [1] -1.141309e-13 -3.841662e-05 -6.293743e-05  4.643397e-12  7.320367e-12 -2.765788e-12 -9.481305e-13 -4.125589e-13  1.794120e-13
[10]  3.249845e-12 -3.369616e-11 -2.640821e-11 -1.170974e-11  3.801404e-13  8.483880e-12  9.634959e-12 -4.114042e-12 -2.496847e-11
[19] -3.993250e-12 -5.382361e-12 -1.030642e-11  1.278977e-13  6.455281e-12 -4.361409e-05 -2.280842e-12 -3.716393e-05 -1.524150e-10
[28] -3.716393e-05 -2.188028e-12 -3.716393e-05 -8.494538e-12 -6.433230e-05 -4.732925e-11 -6.433230e-05 -2.837908e-11 -6.433230e-05
[37] -8.242296e-12 -1.193712e-12 -4.510525e-11 -1.203038e-12 -2.599876e-11 -1.390443e-12  8.312782e-10

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   3.239324e+00  2.803062e-06 -1.026331e-07 -4.655971e-03 -9.861335e-03  2.215723e-03  4.674075e-04  1.018885e-02  9.897251e-02
   2.803062e-06  3.847401e-05  5.962317e-10 -2.055206e-05 -2.561212e-06  7.009368e-06  5.529732e-07  5.475692e-06  1.561510e-05
  -1.026331e-07  5.962317e-10  6.293238e-05  2.308040e-06 -2.817846e-06 -3.740611e-06  2.272228e-07  4.871991e-06 -1.923879e-08
  -4.655971e-03 -2.055206e-05  2.308040e-06  3.319261e+00  2.982327e-01  4.989480e-01 -1.209624e-01  2.421874e-01 -4.852528e-02
  -9.861335e-03 -2.561212e-06 -2.817846e-06  2.982327e-01  2.257936e+00  6.166996e-01 -1.952613e-02  1.502908e-01  1.930068e-02
   2.215723e-03  7.009368e-06 -3.740611e-06  4.989480e-01  6.166996e-01  9.645158e-01  1.835863e-02 -1.001684e-03 -7.094371e-03
   4.674075e-04  5.529732e-07  2.272228e-07 -1.209624e-01 -1.952613e-02  1.835863e-02  1.313680e+00  5.163769e-01 -2.089642e-02
   1.018885e-02  5.475692e-06  4.871991e-06  2.421874e-01  1.502908e-01 -1.001684e-03  5.163769e-01  2.276831e+00 -2.332673e-02
   9.897251e-02  1.561510e-05 -1.923879e-08 -4.852528e-02  1.930068e-02 -7.094371e-03 -2.089642e-02 -2.332673e-02  1.074505e+00
   1.166017e-02 -1.688647e-05  4.682951e-07 -9.901070e-02  3.460362e-02 -1.274637e-02 -1.216766e-02 -1.635226e-02  3.118834e-01
   2.718922e-03  2.027371e-05 -6.222980e-06 -1.059420e-01 -2.342412e-01 -5.845164e-02 -5.945603e-03  1.450436e-01 -2.179636e-02
  -1.330583e-03  1.740423e-05 -1.284109e-05  1.858594e-02 -3.025115e-01  2.347332e-02 -1.386949e-02  1.245454e-01 -8.154970e-03
   2.034268e-02  1.506556e-05  1.411397e-05  1.428543e-01  3.612044e-02  2.906467e-01  1.365251e-02  1.078804e-01  2.973563e-02
   7.999936e-03 -1.422714e-05 -1.230792e-06  3.448183e-02  8.722206e-02  7.191428e-02 -1.389344e-03  5.877991e-03 -1.080819e-02
   7.855587e-03  2.162046e-05 -9.116873e-07  7.716402e-02  9.254695e-02  2.333290e-01  8.283178e-04 -1.894889e-02  1.254596e-02
  -1.400190e-02  3.745601e-05  1.019634e-05  3.282320e-02  3.265219e-03 -8.068774e-02  1.907175e-02 -1.676053e-01  1.798975e-02
   9.642688e-03  2.224439e-05  1.119093e-06 -6.071376e-02  6.617266e-02  6.670798e-02  1.706888e-02  6.069855e-02 -1.649276e-02
   1.818988e-04  1.174916e-05 -2.701533e-06 -9.309382e-03  1.054132e-01  3.295768e-02 -5.993919e-03 -4.630091e-03 -4.789433e-04
   1.546280e-03  6.194585e-06  4.539471e-06  1.733448e-01  9.707207e-02  8.584662e-03  3.171094e-02 -1.668040e-01  1.750513e-02
   3.682494e-03 -1.084039e-05 -5.422156e-07  5.353464e-02  3.134386e-02  6.493362e-02 -6.177510e-04  5.992361e-03 -1.622999e-02
  -2.343230e-05  3.066585e-06  1.331614e-07 -5.768927e-03  9.865061e-03 -2.067698e-03 -5.174763e-04 -3.479095e-03  2.190981e-03
  -6.335387e-03  2.337854e-05  5.269652e-06  4.808203e-02  1.351251e-01  2.859955e-01  1.828289e-02 -7.824412e-02  1.439726e-02
   1.040901e-02  3.745908e-05  3.801050e-07 -6.311519e-02 -3.838522e-03  5.339741e-03 -1.176498e-02  1.824831e-02  3.139731e-04
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
   1.166017e-02  2.718922e-03 -1.330583e-03  2.034268e-02  7.999936e-03  7.855587e-03 -1.400190e-02  9.642688e-03  1.818988e-04
  -1.688647e-05  2.027371e-05  1.740423e-05  1.506556e-05 -1.422714e-05  2.162046e-05  3.745601e-05  2.224439e-05  1.174916e-05
   4.682951e-07 -6.222980e-06 -1.284109e-05  1.411397e-05 -1.230792e-06 -9.116873e-07  1.019634e-05  1.119093e-06 -2.701533e-06
  -9.901070e-02 -1.059420e-01  1.858594e-02  1.428543e-01  3.448183e-02  7.716402e-02  3.282320e-02 -6.071376e-02 -9.309382e-03
   3.460362e-02 -2.342412e-01 -3.025115e-01  3.612044e-02  8.722206e-02  9.254695e-02  3.265219e-03  6.617266e-02  1.054132e-01
  -1.274637e-02 -5.845164e-02  2.347332e-02  2.906467e-01  7.191428e-02  2.333290e-01 -8.068774e-02  6.670798e-02  3.295768e-02
  -1.216766e-02 -5.945603e-03 -1.386949e-02  1.365251e-02 -1.389344e-03  8.283178e-04  1.907175e-02  1.706888e-02 -5.993919e-03
  -1.635226e-02  1.450436e-01  1.245454e-01  1.078804e-01  5.877991e-03 -1.894889e-02 -1.676053e-01  6.069855e-02 -4.630091e-03
   3.118834e-01 -2.179636e-02 -8.154970e-03  2.973563e-02 -1.080819e-02  1.254596e-02  1.798975e-02 -1.649276e-02 -4.789433e-04
   1.888598e+00  1.491555e-03 -2.101002e-03  8.206079e-02 -1.771091e-02  6.615848e-03 -1.398332e-01 -6.559100e-02 -2.694208e-02
   1.491555e-03  1.018173e+00  1.195982e-01  1.562365e-02  2.320134e-02  5.011303e-02 -7.032886e-03  6.978992e-04  5.569214e-02
  -2.101002e-03  1.195982e-01  8.367502e-01  8.754933e-02 -7.321984e-02  3.325485e-02 -3.382396e-02  7.623568e-02  6.867418e-03
   8.206079e-02  1.562365e-02  8.754933e-02  6.806033e+01  5.513000e+00  5.270563e+00 -3.671312e+00  9.943905e-01  1.056249e+00
  -1.771091e-02  2.320134e-02 -7.321984e-02  5.513000e+00  5.418338e+00  1.436004e+00 -2.431270e+00  2.771824e-01  7.497548e-01
   6.615848e-03  5.011303e-02  3.325485e-02  5.270563e+00  1.436004e+00  7.891373e+00 -5.043487e-01  7.460292e-01  8.166589e-01
  -1.398332e-01 -7.032886e-03 -3.382396e-02 -3.671312e+00 -2.431270e+00 -5.043487e-01  3.645773e+01  6.884865e-01 -6.045563e-02
  -6.559100e-02  6.978992e-04  7.623568e-02  9.943905e-01  2.771824e-01  7.460292e-01  6.884865e-01  1.111065e+01  1.498566e-01
  -2.694208e-02  5.569214e-02  6.867418e-03  1.056249e+00  7.497548e-01  8.166589e-01 -6.045563e-02  1.498566e-01  3.087727e+00
  -2.087469e-02 -1.185052e-02  1.249112e-01 -3.596606e+00 -7.318534e-01 -3.570245e+00  7.430528e+00  1.325755e-01 -2.556191e-01
  -1.823780e-02  6.293489e-02  1.054980e-01  6.318349e-01  2.565299e-01 -2.175935e-02 -1.150282e-01  2.336562e+00 -1.280794e-01
  -7.144739e-03 -1.939257e-03  2.172193e-03  3.938966e-02  9.195528e-02  1.561159e-02  1.887339e-02  2.606203e-02  2.988804e-01
  -1.057301e-02  6.244597e-02  1.552664e-01 -6.767114e+00 -4.101492e-01 -5.039288e-01  1.167780e+01  1.680130e-01  3.763661e-02
  -7.282526e-02  8.982698e-02  1.175569e-01  1.650057e+00  8.955661e-03  1.634812e-01 -3.418297e-01  3.363876e+00 -8.824409e-02
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   1.546280e-03  3.682494e-03 -2.343230e-05 -6.335387e-03  1.040901e-02  3.620778e-08 -9.685577e-02  3.005054e-08 -1.589186e-04
   6.194585e-06 -1.084039e-05  3.066585e-06  2.337854e-05  3.745908e-05  3.078347e-10 -2.512951e-06  5.927509e-10  7.474214e-06
   4.539471e-06 -5.422156e-07  1.331614e-07  5.269652e-06  3.801050e-07  4.077985e-11 -2.173012e-07 -1.346886e-10 -1.056814e-05
   1.733448e-01  5.353464e-02 -5.768927e-03  4.808203e-02 -6.311519e-02 -1.943278e-07 -7.259500e-02 -4.831440e-07 -6.792238e-03
   9.707207e-02  3.134386e-02  9.865061e-03  1.351251e-01 -3.838522e-03  3.407026e-06  1.467679e-02 -1.024895e-07 -2.886798e-02
   8.584662e-03  6.493362e-02 -2.067698e-03  2.859955e-01  5.339741e-03  1.941815e-07 -3.541286e-02  4.165546e-06 -2.757134e-02
   3.171094e-02 -6.177510e-04 -5.174763e-04  1.828289e-02 -1.176498e-02 -6.090496e-08  1.812157e-02 -2.225425e-07 -2.174313e-04
  -1.668040e-01  5.992361e-03 -3.479095e-03 -7.824412e-02  1.824831e-02  2.020301e-07 -3.267900e-02 -2.445570e-07  1.063137e-02
   1.750513e-02 -1.622999e-02  2.190981e-03  1.439726e-02  3.139731e-04  2.863060e-07 -3.256057e-02 -2.491455e-07  2.965075e-03
  -2.087469e-02 -1.823780e-02 -7.144739e-03 -1.057301e-02 -7.282526e-02 -3.638717e-07 -7.749177e-03 -7.781771e-07 -2.420821e-02
  -1.185052e-02  6.293489e-02 -1.939257e-03  6.244597e-02  8.982698e-02  2.205705e-06  4.620239e-02  5.451343e-07  2.027159e-02
   1.249112e-01  1.054980e-01  2.172193e-03  1.552664e-01  1.175569e-01 -5.127111e-08  2.261266e-02 -2.551712e-06 -6.151003e-03
  -3.596606e+00  6.318349e-01  3.938966e-02 -6.767114e+00  1.650057e+00  2.838380e-05 -4.554702e-04  7.605095e-06 -7.899810e-02
  -7.318534e-01  2.565299e-01  9.195528e-02 -4.101492e-01  8.955661e-03  1.050621e-05 -4.504978e-02 -8.846752e-07  1.518926e-02
  -3.570245e+00 -2.175935e-02  1.561159e-02 -5.039288e-01  1.634812e-01  9.584520e-06 -4.031859e-03 -3.819007e-06 -1.015480e-01
   7.430528e+00 -1.150282e-01  1.887339e-02  1.167780e+01 -3.418297e-01 -9.637132e-07 -7.433937e-02 -1.496415e-05 -1.478104e-01
   1.325755e-01  2.336562e+00  2.606203e-02  1.680130e-01  3.363876e+00  2.847775e-06  7.532108e-02 -2.733200e-06  5.136703e-02
  -2.556191e-01 -1.280794e-01  2.988804e-01  3.763661e-02 -8.824409e-02  3.021544e-05  8.480762e-03  5.743898e-06  4.299691e-01
   2.844958e+01  4.151879e-01 -4.623409e-02  1.318504e+01  2.203544e-03 -2.592027e-06 -6.063706e-02  1.286188e-05 -5.139720e-02
   4.151879e-01  3.201967e+00 -1.214434e-02  7.204437e-02  1.586393e+00 -6.816009e-07 -1.951910e-02 -5.250567e-06  9.235259e-02
  -4.623409e-02 -1.214434e-02  1.863646e-01  2.458068e-02 -1.979537e-02  9.433006e-06  2.865653e-03 -2.255914e-06  9.504020e-02
   1.318504e+01  7.204437e-02  2.458068e-02  2.421742e+01  3.993770e-01  2.284533e-06 -7.676435e-02 -2.040872e-06 -6.330361e-03
   2.203544e-03  1.586393e+00 -1.979537e-02  3.993770e-01  5.685375e+00 -2.363626e-06  5.663015e-02 -3.174914e-06 -9.705308e-02
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
   3.005054e-08  2.589768e-04  3.005054e-08 -2.399594e-02 -5.370128e-08  7.591378e-03 -5.370128e-08  4.014155e-03 -5.370128e-08
   5.927509e-10  7.005194e-06  5.927509e-10 -1.095698e-05  8.994257e-10  4.988482e-06  8.994257e-10  7.447891e-05  8.994257e-10
  -1.346886e-10 -1.650155e-07 -1.346886e-10 -7.788173e-08 -2.141209e-10 -1.517430e-05 -2.141209e-10 -3.819572e-06 -2.141209e-10
  -4.831440e-07 -5.638543e-03 -4.831440e-07  1.071179e-01 -6.547804e-07 -6.072059e-02 -6.547804e-07 -3.845532e-02 -6.547804e-07
  -1.024895e-07 -1.410026e-03 -1.024895e-07  6.099236e-02 -1.747208e-06  1.015925e-02 -1.747208e-06 -1.665329e-02 -1.747208e-06
   4.165546e-06  6.014370e-03  4.165546e-06  2.658481e-01  3.569838e-06 -5.593690e-02  3.569838e-06  8.406758e-02  3.569838e-06
  -2.225425e-07  8.338413e-04 -2.225425e-07  8.683306e-03 -1.923825e-07 -3.296350e-03 -1.923825e-07  4.250631e-02 -1.923825e-07
  -2.445570e-07  3.670315e-03 -2.445570e-07  4.390346e-03  9.283889e-07  5.471059e-02  9.283889e-07  9.096991e-03  9.283889e-07
  -2.491455e-07 -3.029114e-03 -2.491455e-07 -2.518079e-02  5.240570e-07  7.251065e-03  5.240570e-07 -4.513590e-02  5.240570e-07
  -7.781771e-07 -4.495596e-03 -7.781771e-07 -1.360236e-02 -2.977474e-06 -4.228874e-02 -2.977474e-06 -6.824874e-02 -2.977474e-06
   5.451343e-07 -1.617003e-03  5.451343e-07 -1.326625e-02  4.268984e-06  7.554551e-02  4.268984e-06  7.760056e-02  4.268984e-06
  -2.551712e-06  2.522864e-03 -2.551712e-06  9.512365e-02  2.983049e-06  6.295328e-02  2.983049e-06  1.759578e-01  2.983049e-06
   7.605095e-06 -9.827437e-03  7.605095e-06  2.444763e-01 -4.213360e-06  5.171391e-01 -4.213360e-06 -4.513175e-01 -4.213360e-06
  -8.846752e-07  6.059504e-03 -8.846752e-07  2.956199e-02 -2.772122e-05  3.575002e-01 -2.772122e-05 -2.914186e-01 -2.772122e-05
  -3.819007e-06  4.327357e-03 -3.819007e-06 -4.756851e-02  2.383848e-05 -7.572409e-03  2.383848e-05  2.384651e-01  2.383848e-05
  -1.496415e-05 -1.252884e-02 -1.496415e-05 -2.000054e-01 -5.655520e-05  4.145791e-01 -5.655520e-05 -4.082250e-01 -5.655520e-05
  -2.733200e-06 -1.185679e-02 -2.733200e-06 -6.690333e-02 -8.812674e-06 -5.870370e-02 -8.812674e-06 -1.209799e+00 -8.812674e-06
   5.743898e-06  3.133858e-03  5.743898e-06  8.844048e-02  9.103638e-06  6.246060e-01  9.103638e-06  1.282953e-01  9.103638e-06
   1.286188e-05  1.278736e-02  1.286188e-05 -2.921554e-01 -1.071737e-05 -2.702772e-01 -1.071737e-05  2.915477e-01 -1.071737e-05
  -5.250567e-06 -3.341871e-03 -5.250567e-06 -2.844092e-02  3.894293e-06 -2.869911e-01  3.894293e-06 -2.329327e-01  3.894293e-06
  -2.255914e-06 -8.063697e-04 -2.255914e-06  3.505500e-02 -1.140261e-05  1.300857e-01 -1.140261e-05  2.433661e-02 -1.140261e-05
  -2.040872e-06 -1.295330e-02 -2.040872e-06 -2.091751e-01 -1.673767e-05  6.633227e-02 -1.673767e-05  6.797890e-02 -1.673767e-05
  -3.174914e-06 -7.608014e-03 -3.174914e-06 -3.901309e-02  1.862357e-05 -2.931420e-01  1.862357e-05 -1.486983e-01  1.862357e-05
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]         [,43]
   5.148202e-03  2.227418e-03 -3.520200e-03  2.227418e-03 -2.896279e-03  2.227418e-03 -2.928429e+00
  -2.007979e-05  2.584984e-06  3.337298e-05  2.584984e-06 -2.199942e-05  2.584984e-06 -2.481696e-04
  -1.547831e-06 -1.391614e-06 -6.480514e-06 -1.391614e-06 -3.158484e-07 -1.391614e-06 -2.791521e-05
   3.330571e-01  9.366132e-04  9.359706e-03  9.366132e-04  2.154394e-02  9.366132e-04 -1.260142e+01
   8.137558e-02  8.994266e-03  2.307253e-01  8.994266e-03 -1.004802e-01  8.994266e-03 -9.395528e+00
  -1.587983e-01  3.839453e-03  1.617428e-02  3.839453e-03  3.931150e-01  3.839453e-03 -9.558610e+00
  -5.640048e-03  9.295586e-04 -8.900323e-03  9.295586e-04 -4.664057e-03  9.295586e-04 -2.106781e+00
  -6.229886e-02 -4.133885e-03  3.383276e-02 -4.133885e-03  6.653201e-02 -4.133885e-03 -3.451444e+00
  -8.170712e-02  4.813851e-04  2.330849e-02  4.813851e-04  1.740057e-02  4.813851e-04 -2.488524e+00
   7.111882e-02 -9.192905e-03 -5.669800e-02 -9.192905e-03  2.471772e-02 -9.192905e-03 -3.415491e+00
   9.235468e-02  1.158516e-04  7.365949e-02  1.158516e-04  6.458454e-02  1.158516e-04 -3.126158e+00
   2.162964e-02 -4.297287e-03 -1.322587e-01 -4.297287e-03  1.315309e-01 -4.297287e-03 -3.749263e+00
   1.026749e+00  1.253306e-01  1.227346e+00  1.253306e-01  2.967928e-01  1.253306e-01 -9.927052e+01
   1.344217e-01 -4.758047e-02  1.065684e+00 -4.758047e-02  8.218931e-02 -4.758047e-02 -2.733273e+01
   1.034430e+00  5.055142e-02  2.968229e-01  5.055142e-02  5.457636e-01  5.055142e-02 -2.639402e+01
   5.826416e-01 -1.074587e-01  2.061789e-01 -1.074587e-01 -4.650944e-01 -1.074587e-01 -8.045167e+01
   1.457277e-01  6.749625e-02  4.043666e-01  6.749625e-02 -8.500097e-01  6.749625e-02 -4.508928e+01
   7.036980e-02 -3.700545e-02  1.642709e+00 -3.700545e-02  9.810573e-02 -3.700545e-02 -1.742303e+01
   6.568612e-01 -3.725213e-02 -4.530174e-01 -3.725213e-02 -5.731572e-01 -3.725213e-02 -6.895053e+01
   4.291094e-02  8.098445e-03  1.816126e-01  8.098445e-03 -1.972566e-02  8.098445e-03 -1.845467e+01
   4.224093e-02  2.712163e-03  3.576262e-01  2.712163e-03 -3.615675e-02  2.712163e-03 -2.736019e+00
   7.746138e-01 -3.269082e-02 -3.546063e-02 -3.269082e-02 -2.544120e-01 -3.269082e-02 -6.038322e+01
   1.527755e-01 -2.370720e-02 -2.402400e-01 -2.370720e-02 -2.841807e-01 -2.370720e-02 -2.666874e+01
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  10417 / 10417 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value   
s(measurement.no)                            9.00e+00 6.86e+00    1.02   0.930   
s(f0Zscore2)                                 9.00e+00 1.00e+00    1.02   0.870   
s(durationZscore2)                           2.90e+01 1.00e+00    1.03   0.985   
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 6.41e+01    0.94   0.005 **
ti(measurement.no,durationZscore2)           4.40e+01 1.21e+01    0.98   0.135   
ti(measurement.no,f0Zscore2)                 2.80e+01 1.28e+01    1.02   0.895   
ti(f0Zscore2,durationZscore2)                7.70e+01 1.48e+01    1.02   0.805   
s(word)                                      3.19e+02 1.99e+02      NA      NA   
s(wordPos)                                   4.66e+02 5.47e+01      NA      NA   
s(wordLeftRightTone)                         4.43e+02 5.28e+01      NA      NA   
s(measurement.no,wordPos)                    4.66e+02 1.61e+02      NA      NA   
s(f0Zscore2,wordPos)                         4.66e+02 9.02e+01      NA      NA   
s(durationZscore2,wordPos)                   4.66e+02 3.48e+01      NA      NA   
s(measurement.no,wordLeftRightTone)          4.43e+02 1.38e+02      NA      NA   
s(f0Zscore2,wordLeftRightTone)               4.43e+02 3.69e+01      NA      NA   
s(durationZscore2,wordLeftRightTone)         4.43e+02 5.47e+00      NA      NA   
s(measurement.no,word)                       3.19e+02 1.21e+02      NA      NA   
s(f0Zscore2,word)                            3.19e+02 5.33e+01      NA      NA   
s(durationZscore2,word)                      3.19e+02 6.72e-04      NA      NA   
s(measurement.no,speaker)                    1.00e+02 2.69e+01    1.02   0.920   
s(durationZscore2,speaker)                   3.00e+02 2.12e+01    1.03   0.995   
s(f0Zscore2,speaker)                         1.00e+02 2.27e+00    1.02   0.855   
s(measurement.no,speakerPos)                 3.00e+02 3.97e+01    1.02   0.935   
s(durationZscore2,speakerPos)                9.00e+02 8.71e+01    1.03   0.995   
s(f0Zscore2,speakerPos)                      3.00e+02 7.66e+01    1.02   0.840   
s(measurement.no,speakerLeftRightTone)       6.00e+02 1.54e+02    1.02   0.920   
s(durationZscore2,speakerLeftRightTone)      1.80e+03 1.01e+02    1.03   0.995   
s(f0Zscore2,speakerLeftRightTone)            6.00e+02 1.19e+02    1.02   0.850   
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# 3D plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.680380067515762. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0990769355270091. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au),
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.530017553918074. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

# 3D plotting
 png("pred2b-7.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.25))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.680380067515762. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-8.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.mas$durationZscore2, c(0.5))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.0990769355270091. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

png("pred2b-9.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2g, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.mas$durationZscore2, c(0.75))),
        ylim = quantile(data.au.mas$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75au * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0au + global_mean0au),
        transform = function(f2Zscore2) f2Zscore2 * global_sd2au + global_mean2au, 
        zlim = c(1000, 1400),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.253736 to 1.355745. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.530017553918074. 
    * word : factor; set to the value(s): 到. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 到 Med. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 到 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0005. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0004 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0005 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1000, 1400), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

5.10 Model 2H: /au/ female, F2 as output

system.time(gamm.model2h.noAR <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
                          
                          
 
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores)
)

#saveRDS(gamm.model2h.noAR, paste("Gamm_model2h_noAR.rds"))

gamm.model2h.noAR <- 
  readRDS("Gamm_model2h_noAR.rds")

r.gamm.model2h <- start_value_rho(gamm.model2h.noAR)

# Auto-regressive model

system.time(gamm.model2h <- bam(f2Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1), 
 
                        
                    
                      data=data.au.fem, method="fREML", discrete = TRUE, nthreads = ncores, rho = r.gamm.model2h, AR.start = data.au.fem$start)
)

#saveRDS(gamm.model2h, paste("Gamm_model2h.rds"))

gamm.model2h <- 
  readRDS("Gamm_model2h.rds")

summary(gamm.model2h, re.test = FALSE)

Family: gaussian 
Link function: identity 

Formula:
f2Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Parametric coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept)  0.02958    0.06649   0.445    0.656

Approximate significance of smooth terms:
                                                edf Ref.df      F  p-value    
s(measurement.no)                             7.016  7.459 46.999  < 2e-16 ***
s(f0Zscore2)                                  1.000  1.000 18.340 1.88e-05 ***
s(durationZscore2)                            2.747  2.940  1.431   0.2266    
ti(measurement.no,f0Zscore2,durationZscore2) 44.074 64.013  1.320   0.0448 *  
ti(measurement.no,durationZscore2)            7.952 10.232  3.701 5.19e-05 ***
ti(measurement.no,f0Zscore2)                 10.475 13.895  3.465 1.27e-05 ***
ti(f0Zscore2,durationZscore2)                10.158 13.317  1.485   0.1155    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

R-sq.(adj) =   0.75   Deviance explained = 79.1%
fREML = 7659.6  Scale est. = 0.22268   n = 9332

gam.check(gamm.model2h)

Method: fREML   Optimizer: perf chol
$grad
 [1] -3.074910e-08 -3.637654e-05 -1.926216e-06  8.731321e-08  3.519330e-07 -8.500149e-08  1.447793e-08 -7.361485e-08 -1.978111e-08
[10] -1.164051e-07  2.246844e-07 -1.259442e-07 -8.593502e-06 -2.606808e-07 -2.164774e-06  3.054713e-06 -5.065912e-08 -7.266396e-06
[19]  2.538665e-06  5.390557e-07 -1.111610e-06  5.055001e-07  8.653278e-07 -3.596819e-05  1.793014e-08 -5.266526e-05 -9.189334e-05
[28] -5.266526e-05 -5.119083e-05 -5.266526e-05 -7.936636e-07 -1.264994e-07 -8.526895e-06 -1.264992e-07  1.628905e-07 -1.264992e-07
[37]  6.869135e-07  6.390165e-07 -9.524021e-06  6.390161e-07  8.489512e-07  6.390161e-07  1.255051e-04

$hess
           [,1]          [,2]          [,3]          [,4]          [,5]          [,6]          [,7]          [,8]          [,9]
   2.819424e+00  1.975733e-07 -5.044593e-04 -5.631152e-03 -1.513664e-03 -3.909292e-03  2.621727e-03  6.175609e-03  8.126456e-03
   1.975733e-07  3.637545e-05  6.591230e-07  3.903702e-07 -1.670656e-06  4.570402e-07 -3.047568e-07 -7.420139e-08  1.305448e-07
  -5.044593e-04  6.591230e-07  4.846326e-01  1.440748e-02 -1.413005e-03 -2.473398e-02  2.595891e-03  1.236304e-02  8.393302e-04
  -5.631152e-03  3.903702e-07  1.440748e-02  2.054033e+00  4.262546e-01 -3.981307e-01  9.272292e-02  8.893814e-02 -1.253374e-03
  -1.513664e-03 -1.670656e-06 -1.413005e-03  4.262546e-01  1.432052e+00  5.436040e-01 -3.408242e-02 -9.470631e-02 -1.888229e-02
  -3.909292e-03  4.570402e-07 -2.473398e-02 -3.981307e-01  5.436040e-01  1.109710e+00 -1.984474e-03 -3.083361e-02 -4.001879e-02
   2.621727e-03 -3.047568e-07  2.595891e-03  9.272292e-02 -3.408242e-02 -1.984474e-03  1.049566e+00  9.777025e-02  5.894589e-03
   6.175609e-03 -7.420139e-08  1.236304e-02  8.893814e-02 -9.470631e-02 -3.083361e-02  9.777025e-02  1.160625e+00 -2.477564e-02
   8.126456e-03  1.305448e-07  8.393302e-04 -1.253374e-03 -1.888229e-02 -4.001879e-02  5.894589e-03 -2.477564e-02  1.157033e+00
  -2.650389e-03  1.529856e-06 -5.631072e-03 -1.562645e-02 -2.365951e-02 -5.479513e-02  7.998057e-03 -4.738889e-03  3.506961e-01
  -2.485612e-03  1.015333e-06 -4.559181e-02  4.439917e-02 -8.252614e-02 -1.006966e-02 -1.668288e-03  3.382310e-02  2.130685e-03
  -1.262873e-03  1.344682e-06 -9.812693e-03 -1.575377e-03  3.545358e-02 -2.081622e-02 -1.096051e-03  1.904333e-02  1.015311e-02
   1.068930e-02  1.288450e-06  8.551100e-02  1.884414e-03 -2.680559e-04  1.245235e-01  1.065589e-02  1.541097e-02  5.340898e-06
  -7.008830e-04 -9.181646e-07 -4.617555e-03 -1.011517e-03  8.043250e-04 -2.499413e-02 -4.826454e-03 -2.507136e-03  2.420797e-04
   4.768416e-03 -1.940081e-06 -3.117701e-02  2.111386e-02 -3.534657e-02  4.454078e-02 -2.386113e-03  1.205165e-02  1.073177e-03
   6.844733e-04 -3.574772e-06 -8.102747e-02  3.033297e-02 -2.682510e-01 -6.197571e-02  3.365899e-02  1.504440e-01 -1.494562e-02
   3.430893e-03  2.447680e-06  1.483707e-02 -2.498008e-02 -3.057494e-02  1.007059e-02  1.478827e-02  4.395746e-02 -4.920250e-03
   4.350886e-03  1.253294e-06  1.724080e-01 -2.539813e-02 -5.984789e-02 -2.870816e-02 -9.985796e-04 -1.278485e-02 -7.911921e-03
   1.720913e-02 -2.726290e-06 -7.500690e-02  3.284903e-02 -8.329645e-02  9.175296e-03  1.686579e-02  2.378991e-02  2.137647e-02
   6.873664e-03  5.752131e-08 -5.406769e-03  1.369049e-02 -9.555065e-02  3.118958e-02  8.482350e-03 -1.725988e-03  3.512805e-02
  -2.975196e-05  1.057280e-07 -8.102377e-03 -2.891052e-03 -2.301823e-02 -2.095578e-02  2.284902e-04  3.077600e-03 -7.191830e-03
   7.809524e-03  1.022572e-06 -1.246575e-02 -5.332421e-03 -2.493703e-02  1.235061e-01  1.720188e-02  6.599437e-02  2.427246e-02
   1.264045e-02  4.442951e-06  2.283823e-02 -3.326595e-02 -3.505791e-02 -2.382100e-02  4.569288e-03  8.895433e-03  3.444666e-03
          [,10]         [,11]         [,12]         [,13]         [,14]         [,15]         [,16]         [,17]         [,18]
  -2.650389e-03 -2.485612e-03 -1.262873e-03  1.068930e-02 -7.008830e-04  4.768416e-03  6.844733e-04  3.430893e-03  4.350886e-03
   1.529856e-06  1.015333e-06  1.344682e-06  1.288450e-06 -9.181646e-07 -1.940081e-06 -3.574772e-06  2.447680e-06  1.253294e-06
  -5.631072e-03 -4.559181e-02 -9.812693e-03  8.551100e-02 -4.617555e-03 -3.117701e-02 -8.102747e-02  1.483707e-02  1.724080e-01
  -1.562645e-02  4.439917e-02 -1.575377e-03  1.884414e-03 -1.011517e-03  2.111386e-02  3.033297e-02 -2.498008e-02 -2.539813e-02
  -2.365951e-02 -8.252614e-02  3.545358e-02 -2.680559e-04  8.043250e-04 -3.534657e-02 -2.682510e-01 -3.057494e-02 -5.984789e-02
  -5.479513e-02 -1.006966e-02 -2.081622e-02  1.245235e-01 -2.499413e-02  4.454078e-02 -6.197571e-02  1.007059e-02 -2.870816e-02
   7.998057e-03 -1.668288e-03 -1.096051e-03  1.065589e-02 -4.826454e-03 -2.386113e-03  3.365899e-02  1.478827e-02 -9.985796e-04
  -4.738889e-03  3.382310e-02  1.904333e-02  1.541097e-02 -2.507136e-03  1.205165e-02  1.504440e-01  4.395746e-02 -1.278485e-02
   3.506961e-01  2.130685e-03  1.015311e-02  5.340898e-06  2.420797e-04  1.073177e-03 -1.494562e-02 -4.920250e-03 -7.911921e-03
   1.231807e+00 -4.156099e-03  2.782802e-03 -8.552269e-02  1.782634e-02  1.788545e-02  3.168668e-02  4.862505e-03  5.717694e-02
  -4.156099e-03  5.653481e-01  1.627182e-01  5.353121e-02  9.585895e-03  1.838587e-02  5.554201e-02  6.203627e-02 -1.885412e-02
   2.782802e-03  1.627182e-01  3.662550e-01  1.583657e-02 -5.563460e-03  1.454397e-02 -2.733056e-02  3.618136e-02  5.452826e-03
  -8.552269e-02  5.353121e-02  1.583657e-02  9.532817e+01  1.159283e+00  2.729884e+00 -2.582184e+00  9.579803e-01  9.737487e-01
   1.782634e-02  9.585895e-03 -5.563460e-03  1.159283e+00  8.605487e-01  1.376037e-02 -6.646631e-01  5.448329e-02  1.958641e-01
   1.788545e-02  1.838587e-02  1.454397e-02  2.729884e+00  1.376037e-02  4.254433e+00 -2.032106e-01  1.104507e-01  1.488388e-02
   3.168668e-02  5.554201e-02 -2.733056e-02 -2.582184e+00 -6.646631e-01 -2.032106e-01  2.696057e+01  1.797286e-01 -1.265438e-01
   4.862505e-03  6.203627e-02  3.618136e-02  9.579803e-01  5.448329e-02  1.104507e-01  1.797286e-01  2.387628e+00  1.487298e-01
   5.717694e-02 -1.885412e-02  5.452826e-03  9.737487e-01  1.958641e-01  1.488388e-02 -1.265438e-01  1.487298e-01  6.205983e+00
   7.266276e-02  2.513867e-02 -7.133297e-02 -3.571490e+00 -1.397834e-01 -1.891429e+00  3.629879e+00  2.529213e-01 -9.104243e-02
  -1.486265e-02  5.303797e-03 -6.769704e-03  1.359818e+00 -7.354154e-04  5.418453e-01  2.955080e-01  8.245805e-01  1.353841e-01
   6.849000e-03  1.727120e-02 -1.984697e-03  8.569469e-02  3.286764e-03  1.280204e-01 -1.005901e-01 -2.968785e-02  1.772153e-01
   9.383882e-02 -2.243555e-02 -1.090489e-01 -1.062575e+01 -1.333027e-01 -3.232221e-01  9.235006e+00  3.781154e-01 -5.005522e-02
   9.787549e-03  2.152005e-02  3.134963e-02  1.357363e+00 -5.409543e-02 -7.738760e-02  9.202124e-02  7.785550e-01  7.191291e-03
          [,19]         [,20]         [,21]         [,22]         [,23]         [,24]         [,25]         [,26]         [,27]
   1.720913e-02  6.873664e-03 -2.975196e-05  7.809524e-03  1.264045e-02  3.074934e-08  3.588458e-02 -2.508458e-08  4.697637e-08
  -2.726290e-06  5.752131e-08  1.057280e-07  1.022572e-06  4.442951e-06  5.792196e-13  7.052742e-07  1.505323e-11  2.953145e-12
  -7.500690e-02 -5.406769e-03 -8.102377e-03 -1.246575e-02  2.283823e-02  1.926306e-06  1.271737e-02  8.081794e-08  5.762046e-07
   3.284903e-02  1.369049e-02 -2.891052e-03 -5.332421e-03 -3.326595e-02 -8.731664e-08 -2.217246e-02 -5.900148e-07 -2.547591e-07
  -8.329645e-02 -9.555065e-02 -2.301823e-02 -2.493703e-02 -3.505791e-02 -3.519584e-07  4.059962e-02  4.198513e-07  1.373846e-06
   9.175296e-03  3.118958e-02 -2.095578e-02  1.235061e-01 -2.382100e-02  8.500759e-08  1.344333e-02 -2.460753e-07 -2.279234e-07
   1.686579e-02  8.482350e-03  2.284902e-04  1.720188e-02  4.569288e-03 -1.447795e-08 -4.166032e-04 -1.028120e-07 -5.865910e-08
   2.378991e-02 -1.725988e-03  3.077600e-03  6.599437e-02  8.895433e-03  7.361880e-08  1.875136e-02 -3.527128e-07  5.248041e-08
   2.137647e-02  3.512805e-02 -7.191830e-03  2.427246e-02  3.444666e-03  1.978116e-08 -7.246660e-03  4.431183e-07 -7.861414e-08
   7.266276e-02 -1.486265e-02  6.849000e-03  9.383882e-02  9.787549e-03  1.164080e-07 -5.412549e-02  5.610589e-07 -3.194195e-07
   2.513867e-02  5.303797e-03  1.727120e-02 -2.243555e-02  2.152005e-02 -2.246878e-07  1.005856e-02 -8.620661e-09  1.003541e-07
  -7.133297e-02 -6.769704e-03 -1.984697e-03 -1.090489e-01  3.134963e-02  1.259547e-07  3.021037e-02 -9.225252e-08  1.529198e-06
  -3.571490e+00  1.359818e+00  8.569469e-02 -1.062575e+01  1.357363e+00  8.593545e-06 -4.974897e-02 -1.362393e-05 -2.032699e-06
  -1.397834e-01 -7.354154e-04  3.286764e-03 -1.333027e-01 -5.409543e-02  2.607067e-07 -2.045405e-02 -1.601093e-07  5.530161e-08
  -1.891429e+00  5.418453e-01  1.280204e-01 -3.232221e-01 -7.738760e-02  2.164875e-06  2.447375e-02 -2.408077e-06  2.317339e-06
   3.629879e+00  2.955080e-01 -1.005901e-01  9.235006e+00  9.202124e-02 -3.054712e-06 -6.676429e-02 -1.006258e-05  4.909891e-06
   2.529213e-01  8.245805e-01 -2.968785e-02  3.781154e-01  7.785550e-01  5.065245e-08 -2.878445e-02 -4.915216e-07 -4.513529e-06
  -9.104243e-02  1.353841e-01  1.772153e-01 -5.005522e-02  7.191291e-03  7.266795e-06  1.753752e-02 -5.926389e-06  1.265415e-05
   2.332215e+01  3.107489e-01 -1.128063e-01  1.261222e+01  1.277453e-01 -2.538719e-06  8.082005e-02 -1.816275e-06 -6.326317e-07
   3.107489e-01  4.090426e+00  1.370523e-02  5.775857e-01  2.338613e+00 -5.390432e-07 -5.938887e-02  1.875689e-06 -4.537754e-06
  -1.128063e-01  1.370523e-02  4.917012e-01  3.557994e-02 -4.481486e-02  1.111718e-06 -4.973360e-03  1.668509e-06  2.694233e-06
   1.261222e+01  5.775857e-01  3.557994e-02  4.431741e+01  4.530376e-01 -5.054967e-07  2.601711e-02 -2.848169e-06  8.562242e-07
   1.277453e-01  2.338613e+00 -4.481486e-02  4.530376e-01  4.253574e+00 -8.653623e-07 -3.163808e-02  1.111932e-06 -3.017403e-06
          [,28]         [,29]         [,30]         [,31]         [,32]         [,33]         [,34]         [,35]         [,36]
  -2.508458e-08  2.743084e-07 -2.508458e-08 -8.053098e-03 -8.129788e-04  4.001749e-03 -8.129788e-04  2.170559e-02 -8.129788e-04
   1.505323e-11 -2.506059e-10  1.505323e-11  1.698083e-06  2.938664e-07  2.021162e-06  2.938664e-07 -2.420987e-06  2.938664e-07
   8.081794e-08 -5.727225e-08  8.081794e-08  1.425529e-03 -3.689326e-03  1.638905e-02 -3.689326e-03  2.647542e-02 -3.689326e-03
  -5.900148e-07  2.742507e-07 -5.900148e-07 -1.671713e-02 -2.799412e-03 -2.496519e-02 -2.799412e-03  5.294025e-02 -2.799412e-03
   4.198513e-07  9.122447e-07  4.198513e-07 -2.225649e-01 -2.131032e-03  1.226465e-01 -2.131032e-03  6.385028e-02 -2.131032e-03
  -2.460753e-07  1.267931e-07 -2.460753e-07 -1.227609e-01 -4.913595e-03  9.357818e-02 -4.913595e-03  5.718397e-02 -4.913595e-03
  -1.028120e-07 -7.417703e-08 -1.028120e-07  2.719154e-02 -1.606171e-03  9.470349e-03 -1.606171e-03 -5.844918e-03 -1.606171e-03
  -3.527128e-07 -4.213924e-07 -3.527128e-07  6.568632e-02 -3.382337e-03 -9.908717e-03 -3.382337e-03 -5.642824e-04 -3.382337e-03
   4.431183e-07  6.265722e-08  4.431183e-07  2.137583e-02 -9.942360e-05 -8.760373e-03 -9.942360e-05 -1.228065e-02 -9.942360e-05
   5.610589e-07  2.311681e-06  5.610589e-07  1.092460e-01  2.573831e-03 -4.612746e-02  2.573831e-03  2.479807e-02  2.573831e-03
  -8.620661e-09  1.647046e-06 -8.620661e-09 -3.176823e-02  2.516344e-03 -2.499194e-02  2.516344e-03 -4.415464e-02  2.516344e-03
  -9.225252e-08  7.928882e-07 -9.225252e-08  6.863965e-03 -3.452722e-03  4.132484e-02 -3.452722e-03 -1.327237e-02 -3.452722e-03
  -1.362393e-05  6.605262e-06 -1.362393e-05 -1.455880e-02  1.600981e-03  3.369748e-01  1.600981e-03  2.824254e-01  1.600981e-03
  -1.601093e-07  7.238926e-07 -1.601093e-07  6.995670e-02  4.065931e-02 -8.040474e-02  4.065931e-02  1.008048e-01  4.065931e-02
  -2.408077e-06 -1.526777e-06 -2.408077e-06  1.773895e-01 -7.932988e-03  7.663906e-02 -7.932988e-03  5.381333e-02 -7.932988e-03
  -1.006258e-05  7.234513e-07 -1.006258e-05 -6.567712e-02  5.067282e-02 -2.340469e-01  5.067282e-02 -7.920297e-02  5.067282e-02
  -4.915216e-07 -1.434811e-06 -4.915216e-07  2.125414e-01  4.614732e-02 -2.383347e-01  4.614732e-02  1.050468e-01  4.614732e-02
  -5.926389e-06 -8.621177e-07 -5.926389e-06  4.128853e-02  1.902874e-03  7.173894e-01  1.902874e-03 -1.521050e-01  1.902874e-03
  -1.816275e-06  4.717240e-07 -1.816275e-06  1.158878e-01 -2.284823e-02 -2.783243e-01 -2.284823e-02 -1.912606e-01 -2.284823e-02
   1.875689e-06 -3.340943e-06  1.875689e-06  1.494499e-01  4.257272e-02 -1.919136e-01  4.257272e-02 -1.362592e-01  4.257272e-02
   1.668509e-06  7.165533e-07  1.668509e-06  3.421037e-02  8.195610e-03  3.196802e-01  8.195610e-03  2.859951e-02  8.195610e-03
  -2.848169e-06 -2.588466e-06 -2.848169e-06 -2.410011e-01 -1.170660e-02  1.551580e-01 -1.170660e-02  1.573606e-01 -1.170660e-02
   1.111932e-06 -3.944206e-06  1.111932e-06  1.457111e-01  8.870324e-03 -2.419463e-01  8.870324e-03 -1.025036e-01  8.870324e-03
          [,37]         [,38]         [,39]         [,40]         [,41]         [,42]         [,43]
   2.340287e-01  1.228208e-03  1.096829e-03  1.228208e-03  3.049599e-03  1.228208e-03 -3.008122e+00
   3.730144e-06 -7.206708e-08 -5.140740e-07 -7.206708e-08 -7.304814e-06 -7.206708e-08 -1.590408e-05
   4.510049e-03  6.439098e-03  3.800150e-03  6.439097e-03 -1.130751e-02  6.439097e-03 -8.736206e-01
  -2.298055e-01  5.350235e-03 -3.033569e-02  5.350235e-03  9.282911e-02  5.350235e-03 -7.609427e+00
  -5.671756e-02 -8.387322e-03  1.370793e-01 -8.387322e-03  1.233666e-01 -8.387322e-03 -7.682242e+00
  -3.136047e-02 -6.779332e-03 -5.667970e-03 -6.779332e-03 -1.241342e-02 -6.779332e-03 -6.245128e+00
  -2.750198e-02 -2.805922e-03 -6.347909e-04 -2.805922e-03  1.528363e-02 -2.805922e-03 -1.398333e+00
   3.818146e-03 -5.715117e-03 -2.314074e-02 -5.715117e-03  6.060386e-02 -5.715117e-03 -2.077802e+00
  -6.986677e-02  6.579260e-03  1.173925e-02  6.579260e-03 -2.512929e-03  6.579260e-03 -2.012613e+00
   2.325560e-02  9.445433e-03  3.445615e-03  9.445433e-03 -5.200756e-02  9.445433e-03 -2.724800e+00
   4.976281e-02  4.534995e-03 -6.956691e-02  4.534995e-03  1.600365e-01  4.534995e-03 -2.207852e+00
   1.883068e-02 -5.861526e-04 -3.703327e-02 -5.861526e-04  2.711922e-02 -5.861526e-04 -2.371234e+00
   7.006119e-01 -4.929979e-02  3.685132e-01 -4.929979e-02  2.336671e-01 -4.929979e-02 -1.191207e+02
   6.367663e-02 -2.404868e-02 -1.633596e-01 -2.404868e-02 -7.050540e-03 -2.404868e-02 -7.828701e+00
   8.556506e-01 -1.484905e-02  1.215262e-01 -1.484905e-02  1.147344e-01 -1.484905e-02 -1.431609e+01
   4.999794e-01 -1.494489e-02  8.061140e-02 -1.494489e-02 -8.398972e-02 -1.494489e-02 -6.177537e+01
  -8.928093e-02  2.464974e-03  2.408658e-01  2.464974e-03  6.967515e-02  2.464974e-03 -1.772666e+01
   1.184449e-01 -3.120488e-02  1.576090e+00 -3.120488e-02 -8.333710e-03 -3.120488e-02 -2.086273e+01
   2.841656e+00 -1.872214e-01 -2.425895e-01 -1.872214e-01  4.712206e-01 -1.872214e-01 -5.638581e+01
   2.341577e-01  2.640634e-02  6.830314e-02  2.640634e-02  2.414055e-01  2.640634e-02 -2.557334e+01
   4.106050e-02  3.054109e-02  1.492015e-01  3.054109e-02  4.087710e-02  3.054109e-02 -4.275921e+00
   1.736595e+00  3.102809e-03  1.695624e-01  3.102808e-03  6.970199e-03  3.102808e-03 -7.901454e+01
   2.710713e-01  3.752186e-02 -3.349090e-02  3.752186e-02  8.262086e-02  3.752186e-02 -2.211411e+01
 [ getOption("max.print") est atteint -- 20 lignes omises ]

Model rank =  10181 / 10181 

Basis dimension (k) checking results. Low p-value (k-index<1) may
indicate that k is too low, especially if edf is close to k'.

                                                   k'      edf k-index p-value  
s(measurement.no)                            9.00e+00 7.02e+00    0.98   0.070 .
s(f0Zscore2)                                 9.00e+00 1.00e+00    0.99   0.250  
s(durationZscore2)                           2.90e+01 2.75e+00    1.00   0.515  
ti(measurement.no,f0Zscore2,durationZscore2) 3.08e+02 4.41e+01    0.98   0.075 .
ti(measurement.no,durationZscore2)           4.40e+01 7.95e+00    1.01   0.820  
ti(measurement.no,f0Zscore2)                 2.80e+01 1.05e+01    1.02   0.965  
ti(f0Zscore2,durationZscore2)                7.70e+01 1.02e+01    0.99   0.145  
s(word)                                      3.10e+02 2.38e+02      NA      NA  
s(wordPos)                                   4.62e+02 1.57e+01      NA      NA  
s(wordLeftRightTone)                         4.22e+02 2.86e+01      NA      NA  
s(measurement.no,wordPos)                    4.62e+02 1.24e+02      NA      NA  
s(f0Zscore2,wordPos)                         4.62e+02 3.55e+01      NA      NA  
s(durationZscore2,wordPos)                   4.62e+02 4.17e+01      NA      NA  
s(measurement.no,wordLeftRightTone)          4.22e+02 1.13e+02      NA      NA  
s(f0Zscore2,wordLeftRightTone)               4.22e+02 5.11e+01      NA      NA  
s(durationZscore2,wordLeftRightTone)         4.22e+02 8.55e+00      NA      NA  
s(measurement.no,word)                       3.10e+02 1.58e+02      NA      NA  
s(f0Zscore2,word)                            3.10e+02 4.42e+01      NA      NA  
s(durationZscore2,word)                      3.10e+02 3.23e-04      NA      NA  
s(measurement.no,speaker)                    1.00e+02 3.07e+01    0.98   0.075 .
s(durationZscore2,speaker)                   3.00e+02 9.76e-04    1.00   0.565  
s(f0Zscore2,speaker)                         1.00e+02 7.65e-04    0.99   0.265  
s(measurement.no,speakerPos)                 3.00e+02 6.93e+01    0.98   0.115  
s(durationZscore2,speakerPos)                9.00e+02 8.06e+01    1.00   0.555  
s(f0Zscore2,speakerPos)                      3.00e+02 5.73e+01    0.99   0.320  
s(measurement.no,speakerLeftRightTone)       5.80e+02 1.53e+02    0.98   0.090 .
s(durationZscore2,speakerLeftRightTone)      1.74e+03 8.92e+01    1.00   0.535  
s(f0Zscore2,speakerLeftRightTone)            5.80e+02 1.05e+02    0.99   0.330  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# 3D plotting

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.25auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.65512883823603. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.5auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.135002585755248. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))

fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        transform.view = c(function(measurement.no) measurement.no * duration0.75auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf),
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 30 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 30 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.536034051323852. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [31] is not a sub-multiple or multiple of the number of rows [30]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)

axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)

NA
NA
NA

# 3D plotting
 png("pred2b-10.png", width = 6, height = 4, units = "in", res = 1000)
par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.25))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.25auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.65512883823603. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2b-11.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2. = quantile(data.au.fem$durationZscore2, c(0.5))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.5auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf), 
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): -0.135002585755248. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

 png("pred2b-12.png", width = 6, height = 4, units = "in", res = 1000)

par(mfrow=c(1,1), cex=1, mar=c(3,3.5,1.5,3.5), mgp=c(2,0.75,0))
fvisgam(gamm.model2h, view=c("measurement.no","f0Zscore2"),
        cond = list(durationZscore2 = quantile(data.au.fem$durationZscore2, c(0.75))),
        ylim = quantile(data.au.fem$f0Zscore2, c(0.1,0.9), na.rm = T),
        n.grid = 100,
        transform.view = c(function(measurement.no) measurement.no * duration0.75auf * 0.1, function(f0Zscore2) f0Zscore2 * global_sd0auf + global_mean0auf),
        transform = function(f2Zscore2) f2Zscore2 * global_sd2auf + global_mean2auf, 
        zlim = c(1100, 1600),
        hide.label = TRUE, add.color.legend = FALSE, rm.ranef = T, main = "", xlab = "Time (ms)", ylab = "F0 (Hz)",font.lab = 2, cex.lab = 1.3, cex.axis = 1.3, xaxt = "n", color = mapcols_pastel, xaxt = "n")

Summary:
    * measurement.no : numeric predictor; with 100 values ranging from 0.000000 to 10.000000. 
    * f0Zscore2 : numeric predictor; with 100 values ranging from -1.270026 to 1.307646. 
    * durationZscore2 : numeric predictor; set to the value(s): 0.536034051323852. 
    * word : factor; set to the value(s): 报道. (Might be canceled as random effect, check below.) 
    * wordPos : factor; set to the value(s): 报道 Fin. (Might be canceled as random effect, check below.) 
    * wordLeftRightTone : factor; set to the value(s): 报道 L H. (Might be canceled as random effect, check below.) 
    * speaker : factor; set to the value(s): S0133. (Might be canceled as random effect, check below.) 
    * speakerPos : factor; set to the value(s): S0129 Fin. (Might be canceled as random effect, check below.) 
    * speakerLeftRightTone : factor; set to the value(s): S0125 L H. (Might be canceled as random effect, check below.) 
    * NOTE : The following random effects columns are canceled: s(word),s(wordPos),s(wordLeftRightTone),s(measurement.no,wordPos),s(f0Zscore2,wordPos),s(durationZscore2,wordPos),s(measurement.no,wordLeftRightTone),s(f0Zscore2,wordLeftRightTone),s(durationZscore2,wordLeftRightTone),s(measurement.no,word),s(f0Zscore2,word),s(durationZscore2,word),s(measurement.no,speaker),s(durationZscore2,speaker),s(f0Zscore2,speaker),s(measurement.no,speakerPos),s(durationZscore2,speakerPos),s(f0Zscore2,speakerPos),s(measurement.no,speakerLeftRightTone),s(durationZscore2,speakerLeftRightTone),s(f0Zscore2,speakerLeftRightTone)
 
    * Note: Transformation function(s) applied to values of x-axis and / or y-axis.

Avis : data length [101] is not a sub-multiple or multiple of the number of rows [100]

gradientLegend(valRange=c(1100, 1600), length=.5, pos=.75, side=4, n.seg = 1, inside=FALSE,font = 1, cex = 1, color = mapcols_pastel)
axis(1, at=tickvals2, labels=ticknames2, las=2, cex.axis=1)
dev.off()

null device 
          1 

6 Model comparison

We take data.ai.mas as the sample data set, Z scored F1 as the prediction. The purpose of model comparison section is to investigate whether incorporating relevant variables (predictors or random effects) leads to a significant enhancement in model performance.

6.1 Random effects

6.1.1 Testing impact of speaker

With only speaker as the random effect.

system.time(REspeaker <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        
                  
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1),
 
                        
                        
                    
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
) 

utilisateur     système      écoulé 
     71.378       3.443       7.261 

6.1.2 Testing impact of speakerLeftRightTone

Add the cross effect: speaker*LeftRightTone.

system.time(REspeakerLeftRightTone <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        
                  
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                       
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
 
                        
                    
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
) 

utilisateur     système      écoulé 
   1041.011       8.183     117.364 

6.1.3 Testing impact of speakerPos

Add the the cross effect: speaker*Pos.

system.time(REspeakerPos <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1),
                        
                    
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
) 

#saveRDS(REspeakerPos, paste("REspeakerPos.rds"))

REspeakerPos <- 
  readRDS("REspeakerPos.rds")

6.1.4 Testing impact of word as random intercept and random slope

Add the random intercept and random slope of word.

system.time(REword <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
 
 
                        
                    
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)
) 

#saveRDS(REword, paste("REword.rds"))

REword <- 
  readRDS("REword.rds")

6.1.5 Testing impact of pos and leftRightTone on word

Add the cross effect word*pos and word*LeftRightTone.

This is the final full model.

REfullmodel <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr", k = 10) +
                        s(f0Zscore2, bs="cr", k = 10) +
                        s(durationZscore2, bs="cr", k = 30) +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", k = c(5, 8, 12)) +
                        ti(measurement.no, durationZscore2, bs = "cr", k = c(5, 12)) +
                        ti(measurement.no, f0Zscore2, bs = "cr", k = c(5, 8)) +
                        ti(f0Zscore2, durationZscore2, bs = "cr", k = c(8, 12)) +
                        
                        # random effect
                        
                        s(word, bs="re") +
                        s(wordPos, bs="re") +
                        s(wordLeftRightTone, bs="re") +
                        
                        s(wordPos, measurement.no, bs = "re") +
                        s(wordPos, f0Zscore2, bs="re") +
                        s(wordPos, durationZscore2, bs="re") +
                          
                        s(wordLeftRightTone, measurement.no, bs = "re") +
                        s(wordLeftRightTone, f0Zscore2, bs="re") +
                        s(wordLeftRightTone, durationZscore2, bs="re") +
                          
                        s(word, measurement.no, bs = "re") +
                        s(word, f0Zscore2, bs="re") +
                        s(word, durationZscore2, bs="re") +
                        
                        s(measurement.no, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speaker, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speaker, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerPos, bs="fs", xt = list(bs="tp"), k=10, m=1) +
 
                        s(measurement.no, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1) +
                        s(durationZscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=30, m=1) +
                        s(f0Zscore2, speakerLeftRightTone, bs="fs", xt = list(bs="tp"), k=10, m=1),
 
                        
                    
                      data=data.ai.mas, method="fREML", discrete = TRUE, nthreads = ncores)

REfullmodel <- gamm.model2a.noAR

REfullmodel <- 
  readRDS("Gamm_model2a_noAR.rds")

6.1.6 compareML

compareML(REword, REfullmodel)

REword: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(word, measurement.no, 
    bs = "re") + s(word, f0Zscore2, bs = "re") + s(word, durationZscore2, 
    bs = "re") + s(measurement.no, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 30, m = 1) + s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speakerPos, bs = "fs", xt = list(bs = "tp"), 
        k = 10, m = 1) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1)

REfullmodel: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(wordPos, 
    bs = "re") + s(wordLeftRightTone, bs = "re") + s(wordPos, 
    measurement.no, bs = "re") + s(wordPos, f0Zscore2, bs = "re") + 
    s(wordPos, durationZscore2, bs = "re") + s(wordLeftRightTone, 
    measurement.no, bs = "re") + s(wordLeftRightTone, f0Zscore2, 
    bs = "re") + s(wordLeftRightTone, durationZscore2, bs = "re") + 
    s(word, measurement.no, bs = "re") + s(word, f0Zscore2, bs = "re") + 
    s(word, durationZscore2, bs = "re") + s(measurement.no, speaker, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerPos, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerPos, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(measurement.no, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 10, 
    m = 1) + s(durationZscore2, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 30, m = 1) + s(f0Zscore2, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1)

Chi-square test of fREML scores
-----

AIC difference: 1216.69, model REfullmodel has lower AIC.

compareML(REspeakerPos,REword)

REspeakerPos: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speakerPos, bs = "fs", xt = list(bs = "tp"), 
        k = 10, m = 1)

REword: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(word, bs = "re") + s(word, measurement.no, 
    bs = "re") + s(word, f0Zscore2, bs = "re") + s(word, durationZscore2, 
    bs = "re") + s(measurement.no, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(durationZscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 30, m = 1) + s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speakerPos, bs = "fs", xt = list(bs = "tp"), 
        k = 10, m = 1) + s(measurement.no, speakerLeftRightTone, 
    bs = "fs", xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1)

Chi-square test of fREML scores
-----

AIC difference: 2565.29, model REword has lower AIC.

compareML(REspeakerLeftRightTone,REspeakerPos)

REspeakerLeftRightTone: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1)

REspeakerPos: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1) + s(measurement.no, speakerPos, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerPos, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speakerPos, bs = "fs", xt = list(bs = "tp"), 
        k = 10, m = 1)

Chi-square test of fREML scores
-----

AIC difference: 1281.25, model REspeakerPos has lower AIC.

compareML(REspeaker, REspeakerLeftRightTone)

REspeaker: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1)

REspeakerLeftRightTone: f1Zscore2 ~ s(measurement.no, bs = "cr", k = 10) + s(f0Zscore2, 
    bs = "cr", k = 10) + s(durationZscore2, bs = "cr", k = 30) + 
    ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr", 
        k = c(5, 8, 12)) + ti(measurement.no, durationZscore2, 
    bs = "cr", k = c(5, 12)) + ti(measurement.no, f0Zscore2, 
    bs = "cr", k = c(5, 8)) + ti(f0Zscore2, durationZscore2, 
    bs = "cr", k = c(8, 12)) + s(measurement.no, speaker, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speaker, bs = "fs", xt = list(bs = "tp"), k = 30, m = 1) + 
    s(f0Zscore2, speaker, bs = "fs", xt = list(bs = "tp"), k = 10, 
        m = 1) + s(measurement.no, speakerLeftRightTone, bs = "fs", 
    xt = list(bs = "tp"), k = 10, m = 1) + s(durationZscore2, 
    speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), k = 30, 
    m = 1) + s(f0Zscore2, speakerLeftRightTone, bs = "fs", xt = list(bs = "tp"), 
    k = 10, m = 1)

Chi-square test of fREML scores
-----

AIC difference: 3652.24, model REspeakerLeftRightTone has lower AIC.

6.2 Fixed effects

Here we don’t specify k-values since it will cause errors.

6.2.1 removing impact of duration and f0

This is the fundamental structure: only normalized time.

system.time(removedurationandf0 <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr"),
                        
                    
                      data=data.ai.mas, method="ML", discrete = TRUE, nthreads = ncores)
) 

saveRDS(removedurationandf0, paste("removedurationandf0.rds"))

removedurationandf0 <- readRDS("removedurationandf0.rds")

6.2.2 removing impact of duration

system.time(removeduration <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr") +
                        
                        s(f0Zscore2, bs="cr") +
                        
                        # interaction between smooths
                        
                        ti(measurement.no, f0Zscore2, bs = "cr"),
                    
                      data=data.ai.mas, method="ML", discrete = TRUE, nthreads = ncores)
) 

saveRDS(removeduration, paste("removeduration.rds"))

removeduration <- readRDS("removeduration.rds")

6.2.3 removing impact of f0

system.time(removef0 <- bam(f1Zscore2 ~ 
                         # smooth
                        s(measurement.no, bs="cr") +
                        
                        s(durationZscore2, bs="cr") +
                        
                        # interaction between smooths
                        
                        ti(measurement.no, durationZscore2, bs = "cr"),
                    
                      data=data.ai.mas, method="ML", discrete = TRUE, nthreads = ncores)
) 

saveRDS(removef0, paste("removef0.rds"))

removef0 <- readRDS("removef0.rds")

6.2.4 This tests your full model

system.time(fullmodel <- bam(f1Zscore2 ~ 
                        # smooth
                        s(measurement.no, bs="cr") +
                        s(f0Zscore2, bs="cr") +
                        s(durationZscore2, bs="cr") +
                        
                        # interaction between smooths
                        ti(measurement.no, f0Zscore2, durationZscore2, bs = "cr") +
                        ti(measurement.no, durationZscore2, bs = "cr") +
                        ti(measurement.no, f0Zscore2, bs = "cr") +
                        ti(f0Zscore2, durationZscore2, bs = "cr"),
                        
                    
                      data=data.ai.mas, method="ML", discrete = TRUE, nthreads = ncores)
) 

saveRDS(fullmodel, paste("fullmodel.rds"))

fullmodel <- readRDS("fullmodel.rds")

6.2.5 compareML

compareML(removeduration, fullmodel)

removeduration: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(f0Zscore2, bs = "cr") + 
    ti(measurement.no, f0Zscore2, bs = "cr")

fullmodel: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(f0Zscore2, bs = "cr") + 
    s(durationZscore2, bs = "cr") + ti(measurement.no, f0Zscore2, 
    durationZscore2, bs = "cr") + ti(measurement.no, durationZscore2, 
    bs = "cr") + ti(measurement.no, f0Zscore2, bs = "cr") + ti(f0Zscore2, 
    durationZscore2, bs = "cr")

Chi-square test of ML scores
-----

AIC difference: 311.33, model fullmodel has lower AIC.

duration is significant

compareML(removef0, fullmodel)

removef0: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(durationZscore2, 
    bs = "cr") + ti(measurement.no, durationZscore2, bs = "cr")

fullmodel: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(f0Zscore2, bs = "cr") + 
    s(durationZscore2, bs = "cr") + ti(measurement.no, f0Zscore2, 
    durationZscore2, bs = "cr") + ti(measurement.no, durationZscore2, 
    bs = "cr") + ti(measurement.no, f0Zscore2, bs = "cr") + ti(f0Zscore2, 
    durationZscore2, bs = "cr")

Chi-square test of ML scores
-----

AIC difference: 3440.09, model fullmodel has lower AIC.

f0 is significant

compareML(removedurationandf0,removef0)

removedurationandf0: f1Zscore2 ~ s(measurement.no, bs = "cr")

removef0: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(durationZscore2, 
    bs = "cr") + ti(measurement.no, durationZscore2, bs = "cr")

Chi-square test of ML scores
-----

AIC difference: 240.25, model removef0 has lower AIC.

duration is significant

compareML(removedurationandf0,removeduration)

removedurationandf0: f1Zscore2 ~ s(measurement.no, bs = "cr")

removeduration: f1Zscore2 ~ s(measurement.no, bs = "cr") + s(f0Zscore2, bs = "cr") + 
    ti(measurement.no, f0Zscore2, bs = "cr")

Chi-square test of ML scores
-----

AIC difference: 3369.00, model removeduration has lower AIC.

f0 is significant

7 session info

sessionInfo()

R version 4.4.2 (2024-10-31)
Platform: x86_64-pc-linux-gnu
Running under: Rocky Linux 8.10 (Green Obsidian)

Matrix products: default
BLAS:   /opt/R/4.4.2/lib64/R/lib/libRblas.so 
LAPACK: /opt/R/4.4.2/lib64/R/lib/libRlapack.so;  LAPACK version 3.12.0

locale:
 [1] LC_CTYPE=fr_FR.UTF-8       LC_NUMERIC=C               LC_TIME=fr_FR.UTF-8        LC_COLLATE=fr_FR.UTF-8     LC_MONETARY=fr_FR.UTF-8   
 [6] LC_MESSAGES=fr_FR.UTF-8    LC_PAPER=fr_FR.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=fr_FR.UTF-8 LC_IDENTIFICATION=C       

time zone: Europe/Paris
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] colorspace_2.1-1  phonR_1.0-7       itsadug_2.4.1     plotfunctions_1.4 mgcv_1.9-1        ggpubr_0.6.0      scales_1.3.0     
 [8] nlme_3.1-166      lubridate_1.9.3   forcats_1.0.0     stringr_1.5.1     dplyr_1.1.4       purrr_1.0.2       readr_2.1.5      
[15] tidyr_1.3.1       tibble_3.2.1      ggplot2_3.5.1     tidyverse_2.0.0  

loaded via a namespace (and not attached):
 [1] sass_0.4.9        utf8_1.2.4        generics_0.1.3    rstatix_0.7.2     stringi_1.8.4     lattice_0.22-6    hms_1.1.3        
 [8] digest_0.6.37     magrittr_2.0.3    timechange_0.3.0  evaluate_1.0.1    grid_4.4.2        fastmap_1.2.0     jsonlite_1.8.9   
[15] Matrix_1.7-1      backports_1.5.0   Formula_1.2-5     fansi_1.0.6       jquerylib_0.1.4   abind_1.4-8       cli_3.6.3        
[22] rlang_1.1.4       munsell_0.5.1     splines_4.4.2     yaml_2.3.10       withr_3.0.2       cachem_1.1.0      colorRamps_2.3.4 
[29] tools_4.4.2       tzdb_0.4.0        ggsignif_0.6.4    broom_1.0.7       vctrs_0.6.5       R6_2.5.1          lifecycle_1.0.4  
[36] car_3.1-3         pkgconfig_2.0.3   pillar_1.9.0      bslib_0.8.0       gtable_0.3.6      rsconnect_1.3.2   glue_1.8.0       
[43] xfun_0.49         tidyselect_1.2.1  rstudioapi_0.17.1 knitr_1.48        htmltools_0.5.8.1 rmarkdown_2.29    carData_3.0-5    
[50] compiler_4.4.2